Skip to main content

Full text of "Primitive fossil gastropods and their bearing on gastropod classification"

See other formats

\jP Biodiversity 

Smithsonian miscellaneous collections. 

Washington :Smithsonian institution,! 862-1 968. 

V. 117(1953): 
Page(s): Page 1 , Page 2, Page 3, Page 4, Page 5, Page 6, Page 7, Page 8, Page 9, Page 10, 
Page 11, Page 12, Page 13, Page 14, Page 15, Page 16, Page 17, Page 18, Page 19, Page 20, 
Page 21 , Page 22, Page 23, Page 24, Page 25, Page 26, Page 27, Page 28, Page 29, Page 30, 
Page 31 , Page 32, Page 33, Page 34, Page 35, Page 36, Page 37, Page 38, Page 39, Page 40, 
Page 41 , Page 42, Page 43, Page 44, Page 45, Page 46, Page 47, Page 48, Page 49, Page 50, 
Page 51 , Page 52, Page 53, Page 54, Page 55, Page 56 

Contributed by: Smithsonian Institution Libraries 
Sponsored by: Smithsonian 

Generated 22 March 2010 1 1 :30 AM 

This page intentionally left blank. 

Cfiarlejf B* anb iWarp^aux Halcott Btscacfi Jfunb 


Research Associate in Paleontology, U, S. National Miisetim 

(With Two Plates) 



With only one exception that comes readily to mind, the various 
classifications of the class Gastropoda in current use are the work of 
neontologists. The living gastropods are classified on the basis of 
their morphology^ largely the anatomy of the soft parts. The fossil 
forms, or at least the older ones, so far as they belong to genera that 
are now extinct, are given the scantest of notice and are distributed 
in an almost haphazard fashion among the families erected primarily 
for living forms. As neontologists have little familiarity with fossils, 
unless it be the more recent ones, they are not especially struck by 
the resulting incongruities. Of course they fail to take the fullest ad- 
vantage of the information that the older extinct fossil forms can 
furnish as to the early history of the class and its bearing on 
phytogeny. Indeed the inaccuracy of such little knowledge as they 
have of the more ancient fossils is apt to lead them astray. 

That the work of the neontologist is nevertheless of the highest 
importance is too obvious to need comment. He has the entire animal 
available to him, including the soft parts, and in the main he has 
made much of his opportunities. 

The paleontologist, on his part, sufifers from the severe handicap 
that he can never observe directly the soft parts of the forms that 
he studies. In a sense he is forced into the role of a mere concholo- 
gist. Unfortunately, many paleontologists, inadequately trained in 
zoology, surrender with resignation, if not with complacency, to what 
appear to be the necessities of the situation. Nevertheless, it is pos- 
sible to infer from fossil shells somewhat more of the probable gen- 
eral anatomy of the soft parts than is commonly done and these 



inferences, if made with due caution, can be useful. Of course they 
do not stand on the same plane as direct observation ; nevertheless 
to neglect them, or to refuse to give them recognition, however 
guarded, would be unscientific indeed. 

The exception to my original statement that the classification of 
the Gastropoda is largely the work of neontologists is the work of 
Wenz, begun in 1938 and unhappily interrupted by his death soon 
after the close of the second World War (Wenz, 1938-1944). He 
lived to complete only that part dealing with the prosobranchs. Wenz 
was a paleontologist with an excellent training. He acquired some 
familiarity with the older fossil gastropods as a pupil of Prof. Em- 
manuel Kayser and especially of Prof. Ernst Koken, of Tiibingen, 
His field of specialization since his student days was Cenozoic non- 
marine gastropods, a field that did not qualify him particularly for 
the task he undertook. The novelties introduced by Wenz in 1938 
into the classification of the Gastropoda were not in the highest cate- 
gories but at the familial level He made a distinct contribution in 
erecting many families, subfamilies, and superfamilies for extinct 
genera for which there had long been a need, but the inherent diffi- 
culties of working with skeletal material alone and his relative un- 
familiarity with the older marine forms made many of his new fami- 
lies mixtures of incongruous elements, and their placement in the 
higher categories is not always fortunate. 

Perhaps the outstanding contribution of Wenz's work in 1 938 to 
the fundamentals of gastropod classification was his suggestion that 
the isolated, symmetrically paired dorsal muscle scars of Tryblidmm 
(Tryblidiacea) might be a very primitive character suggesting the 
segmentation of the chitons (Wenz, 1938, p. 59) • However, in 1938 
he allowed himself to be influenced by this idea in constructing his 
taxonomic hierarchy only to the extent of erecting a separate super- 
family, Tryblidiacea, for the genera with symmetrically paired dorsal 
muscle scars instead of including them with the superficially similar 
Patellacea, as had been done in effect by previous workers. As is 
generally recognized, the symmetr}^ in the Patellacea is secondary and 
superficial, not primitive. 

Two years later Wenz proposed a more radically revised classi- 
fication of the major categories of the Gastropoda (Wenz, 1940)* 
He recognized a major dichotomy within the Gastropoda (excluding 
the Loricata) between w^hat he regarded as two subclasses, the 
Amphigastropoda (bilaterally symmetrical, primitively orthoneurous, 
with a saucer-shaped, conical, or symmetrically spiral shell) and the 
Prosobranchia (asymmetrical, chiastoneurous, with asymmetrically 


coiled shell). He elaborates somewhat his invaluable earlier views on 
the tryblidians but he does not follow the logic of his position and clas- 
sify them with the chitons. Instead, because of the discovery of multi- 
ple paired dorsal muscle scars in the supposed bellerophont Cyrtonella, 
he classifies the bellerophonts with the tryblidians in a subclasSj the 
Amphigastropoda, This action I do not regard as well taken (Knight, 
ig47b^ p. 264, and appendix to this work). Naef, a neontologist, had 
made a somewhat similar division at an earlier date with the Plano- 
spiralia for the bellerophonts (he was unaware of the probable signifi- 
cance of Tryblidmm and its allies or possibly even of their existence) 
and the Turbospiralia for the asymmetrical groups (Naef, igii, 
p, 159). Naef's Planospiralia, unlike Wenz's Amphigastropoda, was 
looked on as streptoneurotis and, of course^ prosobranch. 

In the interval betw^een the first draft of the present paper and its 
completion, an interval required for the preparation of drawings, a 
significant paper on the aspi dob ranch Gastropoda and their evolution 
appeared. This paper, by the distinguished anatomist and physiolo- 
gist, C- M. Yonge (1947), reports the results of some revealing in- 
vestigations on the anatomy and functioning of the pallial organs of 
some aspidobranchs. Yonge does not stop with the recording of ob- 
servations but proceeds to apply his findings to an interpretation of 
gastropod evolution just as I have done from a different set of obser- 
vations. Both Yonge and I have accepted certain findings and inter- 
pretations from previous workers and to that extent have a common 
background. Hence it is not surprising that there is much basically 
the same in each interpretation. On the whole our acceptance or re- 
jection of the suggestions of previous workers is gratify ingly similar. 
A minor difference is that he regards Wenz's suggestion that the 
tryblidians are pretorsional gastropods only as possible (Yonge, 1947, 
p. 485)- With some rearrangements and dijfferences in emphasis from 
Wenz I accept this as probable. Yonge regards the bellerophonts as 
prosobranchs, just as I do, and thus rejects Wenz's view that they 
were ''primitively orthoneurous/' However, he appears to harbor an 
unexplained and undocumented idea that although they are symmet- 
rical prosobranchs they had a single dorsal and median retractor 
muscle (Yonge, 1947, p, 49^. fig- 3^^)- It is my view that the bel- 
lerophonts are prosobranch gastropods that have undergone torsion 
and have retained a high degree of primitive bilateral symmetry in- 
cluding a single symmetrical pair of retractor muscles attached at the 
distal ends of the columella (Knight, 1947b) • 

Yonge proposes some phylogenies (Yonge, 1947, p. 490, fig. 31a) 
toward which I am compelled to be skeptical I am skeptical of the 


supposed origin of the Neritacea and the pectinibranchs as branches 
arising independently and directly from the bellerophonts. The great 
expansion of the pleurotomarians in the Paleozoic when they over- 
shadowed all other contemporaneous gastropods in diversity of form 
and number of genera and species provided possibilities that cannot 
be neglected. For example, the asymmetrical neritaceans and the 
pectinibranchs may have been derived from asymmetrical ancestors 
such as some of the numerous and varied pleurotomarians. The 
pleurotomarians (Pleurotomariacea) include much besides ''Pleura- 
tomaria/' That the present-day Tkeodoxus and the pectinibranchs 
are not derived from the present-day " Pleurotomaria'^ one can readily 
concede. It is equally unlikely that the present-day Haliotis^ Scis- 
surella^ the fissurellids, or the patellids are correctly derived from the 
present-day *\PIetirotomaria/^ as Yonge seems to imply. I am not even 
prepared to accept Mikadotrodms beyrichi (Hilgendorf ), the species 
from which Yonge derives most of his ideas of pleurotomarian anat- 
omy, as properly referred to the genus Pleurotomaria Sowerby. That 
all these may have had common ancestors more advanced than 
bellerophonts, i. e,, advanced to the pleurotomarian stage or farther, 
seems probable. 

The classifications of neontologists are based actually on com- 
parative anatomy, that is to say on morphology, from which they 
attempt to infer phylogeny, but phytogeny, or descent with change in 
time, is held very much in the background as an ideal only. The pa- 
leontologist alone has spread before him the time sequence, the order 
in which forms appeared in time. This has been called chronogenesis. 
Chronogenesis is not a perfect tool, for the fossil record is far from 
complete and the recognition of phylogenies involves supposed relation- 
ships inferred from imperfectly known morphological criteria. Never- 
theless, it is a useful tool, if used with caution, and is becoming more 
and more useful as our knowledge of the life of the past increases. 

In summary, all classifications are provisional and hypothetical, 
based on inferences from more or less complete observations of vari- 
ous phenomena. Certainly any classification based wholly on neon- 
tological data or with inadequate attention to or understanding of 
paleontological data must be almost as defective in the very nature 
of the case as would be the converse- The present classification is 
offered as one that at least attempts to give full weight to paleon- 
tological data and their bearing on phylogeny. It is admittedly pro- 



Changes in classification. — The principal novelties of the classifica- 
tion given below consist of the following : 

The Polyplacophora are returned to the Gastropoda as one order 
of a subclass, the Isopleura, proposed by Lankester in 1883. The 
order Monoplacophora {Tryblidium and its allies) is added to the 
Polyplacophora, I am allowing the Aplacophora to stand close to the 
Polyplacophora, as do most authors, although without strong con- 
viction. They do not occur as fossils, and paleontology has no 
light to throw on them. They are probably degenerate, not primitive. 
These three orders will make up the Isopleura. 

In the subclass Anisopleura^ also proposed by Lankester In 1883, are 
included as superorders the Prosobranchia, Opisthobranchia, and Pul- 
monata, while to the primitive prosobranch superfamilies Bellero- 
phontacea and Pleurotomariacea of the order Archaeogastropoda is 
added a third, the Macluritacea. Except as affected by the foregoing 
the remaining more advanced archaeogastropod superfamilies are left 
untouched as Wenz left them, not because Wenz's treatment is satis- 
factory but because a complete revision is beyond the scope of this 
paper. Such a revision is well under way, however, and perhaps in the 
not too distant future the results may be published- 

Phylum MolIuBca Cuvier 
Class Gastropoda Cuvier 

Subclass Isopleura Lankester 

Order Monoplacophora Wenz ^ 
Order Polyplacophora Blainville 
Order Aplacophora Jhering 
Subclass Aiiisopleura Lankester 
Superorder Prosobranchia Milne-Edwards 
Order Archaeogastropoda Thiele 

Superfamily Bellerophontacea Ulrich and Scofield ^ 
Superfamily Pleurotomariacea Wenz ^ 
Superfamily Macluritacea Gill ^ 

Other archaeogastropod superfamilies are not considered here, nor 
are the following orders and superorders: 

Order Mesogastropoda Thiele 
Order Neogastropoda Wenz 
Superorder Opisthobranchia Milne-Edwards 
Superorder Pulmonata Cuvier 
Incertae sedis, PelagieUa Matthew^ 1895, and allies* Possibly not 

^ Considered in some detaiL 

2 Only the earlier, more primitive genera and some living ones considered. 


























5i ^^ 






V o 




o 3 


O u 










J»^ J^ J^ ■ 


















.5 i^ 

















Chronogenesis and range in time. — In order to give an over-all 
view of the range in time of the two orders of the Isopleura and the 
three most primitive superfamilies of the anisopleuran prosobranchs, 
a diagram is presented (fig. i). It will be noted that the major 
dichotomy in time (as well as in morphology) is between the Iso- 
pleura and An isopleura in the early Cambrian, at the beginning of 
the fossil record. It will be noted also that two of the three primi- 
tive superfamilies of the prosobranch Archaeogastropoda, the Mac- 
luritacea and Bellerophontacea, have been extinct since Devonian and 
Triassic times, respectively. Only the long-ranging Pleurotomariacea, 
late Upper Cambrian to Recent^ has carried through in several special- 
ized relic families. These give us some clue to the morphology of the 
soft parts and to the physiology and embryology of the primitive 


Technical terms. — As far as possible the use of technical terms 
(other than the formal scientific names of systematic categories) has 
been avoided. With exceptions to be noted the morphological 
terms employed are so much in general use by both paleontologists 
and malacologists that it seems unnecessary to define them. 

The term "hyperstrophic^' is not a new one but experience suggests 
that many have only a hazy idea of its meaning. It refers to that 

,n^ : : ■-. • ^i^ ^ ^ ' '■**• 

Figure i* — Range in time of the more primitive categories. 

For the benefit of the neontologist interested in geologic time in terms of years 
and tinfamiliar with recent work, the following data are arranged from the 
Report of the Committee on the Measurement of Geologic Time of the Division 
of Geology and Geography, National Research Council, for 1949-1950 (p. 18) : 

BeginniTig in 

approximate Approximate 

number of ItiD^th in 

millions of millions o£ 

Period years ago years 

Quaternary * i ^ 

Tertiary -.,.,,. 60 59 

Cretaceous - . . , * ^ .,.«'.*.,*.. . 130 ?^ 

Jurassic * ^-iSS ^S 

Triassic ..,,*,**•,*,,...... 1S5 30 

Permian » . - ,»..,. 2 ro 25 

Carboniferous , * :26s 55 

Devonian .,»», ^ ,.>...... i ^ ^ ^ >.«> = « ^ . > 3^^ 35 

Silurian - 3^0 40 

Ordovician , . . - , . * . , 44^ 80 

Cambrian ».....««•*., 5^<3 ^^ 

Computed probable errors in beginnings : Quaternary ± 50 thousand years, 
Tertiary ± i to 2 million years, Mesozoic ± 5 million years, and Paleozoic ± 10 
million years. 










Figure 2 

Isotrophic coiling. Syminetrically coiled in a plane with the sides mirror 
images of each other. The example is a diagrammatic restoration of a gen- 
eralized bellerophont. Note tlie symmetrical and paired ctenidia, the rectum 
passing through the pericardium and terminating in the anus between the 
ctenidia and close to the slit, and the symmetrical and paired auricles of the 

Asymmetrical coiling, b, Orthostrophic coiling in the pleurotomarians. Al- 
though the shell is asymmetrical and orthostrophic the rectum stifl passes 
through the pericardium and terminates between the ctenidia close to the 
sHt- Many of the organs such as ctenidia, auricles^ etc. are paired. The 
diagram shows a dextral pleurotomarian. No certainly sinistral ones are 
known. c,d, Orthostrophic coiling at an advanced stage of asymmetry. In 
the dextral forms the right ctenidium and auricle are lost, the rectum has 
moved to the right and no longer passes through the pericardium, and the 
slit has disappeared, c shows a sinistral orthostrophic gastropod and d a 
dextral one. Note that the arrangement of the shell and the internal organs 
in each is the mirror image of the other. A tremendous majority of living 
gastropods are dextral orthostrophic, 

Hyperstrophic coiling, e, sinistral and / dextrah Comparing the dextral 
and sinistral hyperstrophic forms with their orthostrophic counter parts^ the 
relative positions of the corresponding internal organs are the same in each 
but the direction of asymmetry of the shell is reversed. In the dextral 
hyperstrophic form the spire protrudes to the left side instead of to the 
right The internal organization is dextral, but the shell if oriented in the 

[legend contmued on opposite page] 


sort of coiling in which the shell is inverted and what appears to be 
the spire is homologous with the base of orthostrophic forms. It is 
as though the normal spire were pushed through, protruding on the 
side that is normally the base and the side that normally has the spire 
resembles a base. The shell resembles superficially a sinistral shell 
but the soft parts are dextral A hyperstrophic sinistral shell re- 
sembles a dextral one but the soft parts are sinistral. 

The term *' orthostrophic'* is employed for the normal coiHng of 
the great majority of asymmetrical gastropods both dextral and sinis- 
tral. The true sinistral gastropod is in all respects a mirror image of 
a dextral gastropod. 

The term ''isostrophic" is introduced as an adjective to describe 
the sort of coiling that is found in many nautiloid and ammonoid 
cephalopoda, and particularly in the bellerophont gastropods. It may 
be exogastric as in the cephalopods or endogastric as in the gastro- 
pods, Isostrophic coiling is symmetrical with the left and right sides 
mirror images of each other. 

Text figure 2 illustrates the different types of coiling described 

The term "Cambrian*' is employed here in the current American 
sense (Howell et ah, 1944, pp. 993-1004) in which beds of Trema- 
docian age are excluded from the Cambrian, Those are placed as 
late Lower Ordovician. It is important that the European reader bear 
this in mind, 

Ilhistrations, — In addition to certain diagrammatic drawings to 
illustrate various points under discussion I have included drawings of 
generalized restorations of a number of characteristic Paleozoic gen- 
era mentioned in the text. Many of these are yet unfamiliar to any 
but specialists and it is hoped that the drawings will be of assistance 
to the general reader in visualiEing what must be unfamiliar genera 
to many. Although these were* made from actual specimens of spe- 
cies, they are restorations intended to illustrate generic characters 
and are not accurate enough to be used for the identification of 

References. — The list of references will be found on pages 55 to 
56, In the text, references to the list are cited in parentheses by 
author and date. Since I published some years ago descriptions and 

custotnary way with the spire upward appears to be sintstraL In this paper 
all illustrated species judged to be hyperstrophic are oriented with the spire 
downward for this brings the aperture to the same side as in a conven- 
tionally oriented orthostrophic shell See plate a on which the Pleuroto- 
mariacea shown are dextral orthostrophic and the Madnritacea are dextral 


figures of the type species of all names of genera based on Paleozoic 
species published before 1938 (Knight, 194^)? no further references 
to such genera will be given here* References to Paleozoic genera pub- 
lished since 1937 appear in the list. For post- Paleozoic genera the 
reader is referred to Wenz, 1938- 1944, which will suffice for many 
purposes. If this rather rare work is unavailable, many works on 
conchology or malacology will do, 


I am indebted to a number of colleagues who have read my manu- 
script at one stage or another of its development. Among these are 
Dr, G, Arthur Cooper, Dr, H. A. Rehder, and R, Tucker Abbott, of 
the U, S. National Museum, and William T. Clench, of Harvard 
University, Perhaps I am indebted most to Prof. Raymond C. 
Moore, of the University of Kansas, and to Dr. L. R. Cox, of the 
British Museum (Natural History). The former has lent encourage- 
ment over the five or six years that my ideas have been developing on 
paper. Both have given the paper critical readings and have furnished 
the stimulus of dissent from some of the views expressed. I alone am 
responsible for departures from the orthodox, 


In order to arrive at hypotheses worthy of attention one must pro- 
ceed from the known to the unknown, or from the better known to 
the less well known. Therefore it will be profitable to consider at 
this point certain selected zoological data, well known, perhaps^ to the 
neontologist but relatively unfamiliar to many paleontologists. Later, 
paleontological data will be considered. 



The chitons are regarded morphologically as the most conservative 
in the basic pattern of their organization of all living gastropods, if 
not of all Uving mollusks. In respect to certain features, the division 
of the shell ^ into eight plates and the musculature to operate them, 
they appear to be highly specialized. Likewise the remarkable shell 
eyes or aesthetes appear to be developed in some genera, possibly in 

^ I am regarding the polyplacophoran shell as homologous with the shells 
of other mollusks* However, it should be noted that at least one recent worker 
regards it as only analogous (Thiele, 1931, p. 2). Perhaps on further study 
this very fundamental difference will be resolved. 


response, as it were, to the loss of the sense organs of the head.^ It 
is these specialized features, the modifications of the primitive basic 
plan, that distinguish them as chitons. The basic plan of organization 
is bilaterally symmetrical in all significant respects. There is a flat, 
creeping foot and well-differentiated head. The head lacks the usual 
sense organs, possibly due to specialized degeneration. The mouth 
is anterior and the anus posterior, the digestive tract passes through 
the pericardium. The gills appear to be true ctenidia and are ar- 
ranged in pairs in a groove between the shell and the upper surface 
of the foot on each side of the body, dominantly in the posterior part 
(Yonge, 1939)^ The numerous paired ctenidia seem to be metameric 
repetitions of a primary pair that lie on each side of the anus and 
just behind the excretory pore (the postrenal gill). The heart is 
dorsal and posterior. The nervous system is not twisted and shares 
the bilateral symmetry of the rest of the body. Strictly speaking 
there is no pallial cavity, but it seems reasonable to regard the pos- 
terior and lateral parts of the groove between the shell and the foot 
that contains the ctenidia and associated organs as strictly homologous 
with the pallial cavity of more advanced gastropods. 

There is no need here to go into a complete morphological descrip- 
tion of the chitons. The features to which I wish to draw special 
attention are the complete bilateral symmetry of all parts and the 
posterior anus^ gills, and heart. It is these features that are regarded 
as primitive and it is contended that it is the modification of these 
features in the ancestral stock from which the chitons were derived 
ihat gave rise to the Anisopleura, modification primarily through 
torsion and progressively greater asymmetry. 

The chitons appear first in the fossil record in late Cambrian time 
and are living today. They were never abundant and for most of the 
time were very rare. They have varied throughout all that vast ex- 
panse of time very little indeed, 


The pleurotomarians ^ are classified in the subclass Anisopleura, 
superorder Prosobranchia, order Archaeogastropoda. They first ap- 

^ Can it be that the microscopic tubules in the shell of Tryblidium reticidatuui 
Lindstrom described and illustrated by Lindstrom in 1884 (p. 56) are the 
tubules of aesthetes instead of some unknown parasitic organism? 

^ The noncommittal vernacular name for this group is employed at this point 
in its broadest sense, as an informal synonym of Pleurotomariacea. The group 
has been treated at one time or another as a genus, as a family, or as a super- 
family, Wenz in 1938 assigned a little over 20Q genera and subgenera, fossil 
and livingj to the Pleurotomariacea* 


pear in the fossil record in late Upper Cambrian time.^ They were 
the most numerous, varied^ and abundant of all gastropods through- 
out succeeding Paleozoic time. They continue in diminishing num- 
bers and variety through the Mesozoic and carry through to the 
present as a few genera, in a few families, the most abundant and 
diversified of which represent two late specializations for rock cling- 
ing (Haliotidae, Fissurellidae), Another living family, composed of 
rare and very tiny forms^ is the Scissurellidae. The fourth family 
of living pleurotomarians, the Pleurotomariidae, is represented in 
present seas by four rarely seen but large and handsome deep-water 
species, of great morphological interest because they are seemingly 
little-changed descendants of early and primitive anisopleuran gastro- 
pods. Although entire specimens are very rare, there have been a 
number of successful dissections with which are associated the names 
of W, H. Dall, E, L. Bouvier and H. Fischer, and of M. F, Wood- 
ward. Dissections have been made also of some of the abundant but 
specialized Haliotidae, Fissurellidae, and Scissurellidae, but the Pleu- 
rotomariidae appear to be less conspicuously specialized for particular 
environments and therefore more significant for the present purpose/ 
This is no place to consider the minor anatomical details of the 
pleurotomarians but certain major features are of importance for our 
purpose. First, in common with all anisopleurans in which the fea- 
tures are not obscured by later developments, all display the effects 
of torsion in that the primitively posterior anus and pallial complex 

^^M^^l^^MI^^—-^^^ U ^ 1 > 1 III 

^ The genera I refer to, four in number, include three with a deep U-shaped 
or V-shaped sinus in the outer lip regarded by most paleontologists as homol- 
ogous to a slit. These are Sinuopea Ulrich, 191 1 (ph 2, fig, 1), Schtzopea Butts, 
1926 (pi. 2, fig. 2), and Dirhacbopea Ulrich and Bridge, 193 1. The fourth, 
Taefiiospira Ulrich and Bridg€| igjij has a moderately deep pleurotomarian sUt 
and a slit band. 

' It will be well here to point out that primitive prosobranch gastropods have 
not a single retractor muscle but a pair of retractor muscles. For example, the 
living representatives of two of the four existing families of the Pleuro- 
tomarianea, the ScissurelUdae and the Haliotidae> have a pair of shell or retrac- 
tor muscles, and a third, the Fissurellidae, has a crescentic muscle accepted as 
compounded from an original pair. In the Bellerophontacea, supposedly the 
immediate forerunners of the pleurotomarians, there is likewise a single pair; 
and in the Neritacea, seemingly an ancient branch from the pleurotomarian stock, 
and the very primitive Macluritidae there is also a pain In the living repre- 
sentatives of the Pleurotomariidae alone, of the supposedly primitive stocks, 
is there a single retractor muscle. This suggests strongly that in this respect 
these have lost one of the primitive muscles and have advanced far toward 
Calliostoma Swainson, 1840, in the Trochidae, to which they may be more 
closely related than to the more primitive pleurotomarians. 


are found in an anterior position above the head as though they had 
been twisted into that position. All have a helicoidally coiled, asym- 
metrical shell at least in late larval stages. But in sjDite of torsional 
asymmetry and the beginnings of lateral asymmetry they retain, as a 
primitive character fully retained in no other group of living ani- 
sopleuran gastropods, paired visceral organs^ including paired ctenidia, 
paired auricles of the heart, paired kidneys, etc. The digestive tract 
passes through the pericardium and the anus discharges between the 
two paired ctenidia* These are primitive characters and they remind 
one strongly of the bilaterally symmetrical pairing of the homologous 
organs in the isopleuran Polyplacophora. They suggest that the Ani- 
sopleura were derived ultimately from bilaterally symmetrical, iso- 
pleuran ancestors. 

Recent and fossil pleurotomarians always, or nearly always, show 
one distinctive shell feature by which they may be recognized almost 
at a glance. This is an emargination in the outer lip of the shell. In 
some of the earliest species it takes the form of a rather deep U-shaped 
or V-shaped sinus. In others the sinus is V-shaped and it may culmi- 
nate in a short slit or notch. Still later appear forms with a deep slit 
and still other modifications, such as a row of tremata, developed in- 
dependently in several genera, or the apical hole in the shell of the 
typical fissurellids. In all living pleurotomarians that have been ex- 
amined the discharge end of the anal tube lies close to the apex of the 
emargination. Yonge has shown from studies of living examples of 
Haliotis Linne, 1758 (Yonge, 1947, p. 449), of the anatomy of a pre- 
served specimen of Mikadotrochus beyrichi (Hilgendorf ) (op. cit., p, 
454), and of the described anatomy of Incisstira lytteUonensis Smith 
(op, cit*, pp, 449-458), as well as of living examples of the more highly 
specialized Fissurellidae, that the respiratory current is created by cilia 
on the filaments of the ctenidia. The water is drawn into the mantle 
cavity above and on both sides of the head. It passes backward beneath 
the ctenidia, impinging on the osphradia enroute, then upward between 
the ctenidial filaments and outward through the slit or its equivalent. 
Close to the inner end of the slit (or row of tremata) is the anus at 
the distal end of the anal tube. The currents, possibly aided by con- 
tractions, carry the faeces out through the slit. This is a highly im- 
portant matter of sanitation for any gastropod with the anus within 
the mantle cavity and directed anteriorly so as to discharge between 
a pair of ctenidia. Obviously such a mechanism would not be impor- 
tant for forms with a posterior anus, such as isopleurans, or for 
those such as the more advanced anisopleurans with an anterior anus 
but with only one ctenidium on the upstream side, as it were, of the 


ciliary currents passing through the mantle cavity, or secondarily 
with a more or less posterior anus as in the opisthobranchs. 


There are inherent technical difficulties in the rearing and studying 
of such extremely small and delicate organisms as the early embryos 
of primitive gastropods. Because of the complexity of the transfor- 
mations and the confusing dififerences in detail from one species to 
another it is difficult to make generalizations in terms that will be 
valid in detail for even the few forms for which much is accurately 
known of the early ontogeny. Furthermore it is difficult to avoid 
attributing to the embryo adult anatomical features which occur only 
as rudiments, often as only a few cells not obviously organized, if 
present at all in the embryo. For example, it is commonly thought 
that in the process of torsion the gastropod becomes so twisted that 
the pallial cavity with the pallial complex including anus, ctenidia, 
kidneys, hearty etc., is translated bodily from a posterior position to 
an anterior position above the head. In effect this is true but in detail 
it is not, for in such primitive genera as Haliotis and Patella Linne, 
1758, for example, the *'proctodaeum and the solid mesoderm rudi- 
ments of the kidneys are the only representatives of the pallial com- 
plex when torsion begins • Even in Vimpams Mont fort, 1810,* in 
w^hich the developmental stages are abbreviated owing to viviparity, 
the single ctenidium and the visceral part of the pleuro-visceral loop 
do not develop until , . . after torsion is complete" (Crofts, 1937, pp. 
262-263). However, in spite of these difficulties a significant series 
of events does occur in a definite order. 

The early trochophore larva has a dorsal shell gland and a stomo- 
daeum (rudimentary mouth) situated immediately below the ciliated 
ring of the velum on the ventral side. The shell, secreted by the shell 
gland, develops from a small disc to a rather deep cup containing the 
dorsal hump. The proctodaeum (rudimentary anus) is moved ven- 
trally toward the stomodaeum. In the process the rudimentary gut; 
Still without open mouth or anus, is bent into a rough U -shape. This 
operation is called flexure and is regarded as distinct from torsion for 
which, however, it lays the foundation. The rudimentary foot appears 
between the stomodaeum and proctodaeum. The pallial cavity appears 
as an invagination posterior to the foot. In the meantime the shell 

s For Paludina, the name employed by Crofts and other embryologists, I am 
substituting the name Viviparus, today regarded as the correct name of the genus 


has continued to grow and, owing to secretion of shell matter more 
rapidly on the posterior margin, it takes on an exogastric roughly 
nautiloid form with the primitive apex directed forward, 

A highly significant organ, the development of which is completed at 
the end of the pretorsional stage, is the single "velum retractor mus- 
cle" first carefully studied by Crofts in Haliotis (Crofts, 1937). The 
muscle before torsion is asymmetrically placed and slightly spiral in 
such a way that its retraction rotates the dorsal hump in a counter^ 
clockwise direction when viewed dorsally. In passing it will be ad- 
vantageous to note that the velum retractor persists through life in 
H alio f is as the small left-hand shell muscle and that the hypertrophied 
right-hand shell muscle, homologous with the single columellar or re- 
tractor muscle in most gastropods, is not at this stage represented by 
a recognizable rudiment of even a single cell It is probable that the 
left-hand retractor muscle in the adults of the more primitive aspido- 
branchs is entirely homologous with that of Haliotis, In more ad- 
vanced types it is lost before maturity. 

In Haliotis torsion begins at about 30 hours after fertilization of 
the ^gg. Crofts (1937, pp* 233-234) reports that the first 90'' of 
torsion takes place in 3 to 6 hours as a result of contraction of the 
"velvum retractor muscle/' The full 180° torsion is not completed 
until 8 or 10 days later and apparently results from differential 

There are curious differences in both the process and time of tor- 
sion as reported by different authors for different species and even 
for the same species. Some of these differences may be caused by 
the difficulties in observing accurately such small and refractory sub- 
jects, but most of them probably reflect actual differences between 
species. Nevertheless there is general agreement on the fact of ISC'" 
torsion at an early embryonic stage. 

The torsion results in the pallial cavity's moving from a posterior 
to an anterior position relative to the foot. Although the organs of 
the pallial cavity have not yet appeared when torsion begins or are 
extremely rudimentary they eventually mature after torsion in an 
anterior position even though their primitive position must have been 
posterior. Likewise the commissures of the visceral nerve complex 
mature after torsion as though they had been crossed to a figure 8 
during the process, although during torsion they were far too short 
and rudimentary to be crossed. Torsion, of course, affects relations 
of the shell to the head and foot so that its apex points to the rear 
of the head instead of forward. 

Before torsion there is some asymmetry in one respect or another, 


in part no doubt anticipatory in nature and chargeable to accelera- 
tion, but after torsion asymmetry develops apace. It is only less 
marked in those forms that develop primitive paired organs than in 
those that develop only one member of the primitive pair, usually 
the definitive left member. The shell is no longer approximately 
bellerophontiform, but coils in a laterally asymmetrical, helicoid spiral 
with the spire pointing backward. That in certain groups the shell 
then becomes symmetrical {Diodora Gray, 1821, for example) or 
that secondary detorsion occurs (opisthobranchs) with a high degree 
of superficial secondary symmetry is irrelevant to our present dis- 
cussion. Nor is it relevant that in a few forms the torsion is clock- 
wise resulting in sinistrality (see fig, 3), 

To recapitulate, the anisopleuran veliger larva is provided with a 
dorsal shell gland ; the gland secretes a shell that grows by marginal 
accretion and soon becomes cuplike ; concurrently the pallial cavity 
is invaginated and the body, w^ith the rudimentary alimentary canal, 
is bent to a U -shape with the procotodaeum within the pallial cavity 
posterior to the stomodaeum and separated from it only by the rudi- 
mentary foot. The U-shaped bending constitutes flexure. The next 
step is torsion by which the dorsal hump with the pallial cavity is 
twisted 180° in a counterclockwise direction (as seen from above) 
relative to the foot, thus laying the foundation for the prosobranch 
and streptoneurous conditions. Next comes, as a separate step, the 
development of lateral asymmetry and the helicoid spire. There are 
in some advanced stocks still further developments, including detor- 
sion which brings about the opisthobranch condition, euthyneury, and 
in extreme cases secondary symmetry of a high order, 



From anatomy. — It is inconceivable that the living anisopleuran 
gastropods, which show torsional and generally lateral asymmetry 
and which are members of the Mollusca, a phylum characterized by 
basic bilateral symmetry, can be at all primitive in respect to those 
features. The most ancient anisopleuran group with living represen- 
tatives, the Pieurotomariacea, appears first in the late Cambrian, 
Living pleurotomarians show vestiges of bilateral symmetry in the 
retention of paired visceral organs along with full torsional asym- 
metry and laterally asymmetrical coiling. The Polyplacophora living 
today are equally as ancient as the pleurotomarians. They are ob- 
viously specialized in respect to the eight -pieced shell, but they retain 

NO. 13 






FiGur.E 3 

Torsion in the embryo of Viviparus viviparus (Linne). It should be noted 
that owing to the viviparity the developmental stages in Vimparus are ab- 
breviated. Hence for this reason and because of the highly diagrammatic 
nature of the drawings (from Naef, 1911, fig. 8, in part) the picture pre- 
sented IS somewhat oversimplified. It is all the more comprehensible for those 
a. Stage where flexure is in progress but torsion not begun. To the left of the 
figure is the ciliated velum, the cup-shaped shell is above and the rudi- 
mentary foot below. The digestive tract is dotted with the mouth below 
and to the left and the anus high and to the right of the figure, 
h. The beginning of torsion. The mantle cavity has appeared and with the 

anus is turned a little to the right of the animal. 
€, Torsion a little more than halfway completed. The anus and mantle cavity 

are now to the right and a little to the front. 
d. Torsion completed. Note that the mantle cavity with the anus is now in 
front and above the head, it fihal position. The shell has become 
belter ophonti form. 



strict bilateral symmetry: They cannot be ancestral to the pletiroto- 
marians, but they very plausibly point the way to that more remote 
ancestor of both chitons and pleurotomarians which must be looked 
for first in Lower or Middle Cambrian rocks unless it became extinct 
before Cambrian time with its record irretrievable. Several very 
distinguished neontologists have speculated as to the probable nature 
of this common ancestor of both and, indeed^ of all the MoUusca. 
The usual conclusion is that it was a mollusk w^ith a single, low, 
conical shell, bilaterally symmetrical in all respects, with the anus and 
pallial complex in the rear, with a differentiated head and a flat creep- 
ing foot. In epitome, it would have the basic bilateral symmetry of 



Scheme of a hypothetical primitive mollusk viewed from the left side, a, amts ; 
Cj g^ cerebral ganglion; /, foot; g, gill, in the pallial cavity; go, gonad; h, 
heart; k, kidney; lac^ labial commissure; m> mouth; pa, mantle; pan, pallial 
nerve; pe, pericardium; pg, pedal ganglion; pig, pleural ganglion; ra, raditia; 
rpo^ renopericardial orifice ; si, stomach ; stg, stoma togastric ganglion ; vg, 
visceral ganglion. (After Pelseneer.) 

the Polyplacophora but with a single shell, as in the Anisopleura, but 
neither coiled nor asymmetrical. Figure 4 shows a reconstruction of 
such a hypothetical ancestral gastropod, a reconstruction based on 
pure deduction before anyone had suspected the possibility that sup- 
posed Cambrian capulids or patellids that we now recognize as the 
Monoplacophora had just about the same anatomy. Figure 5 shows 
a restoration of a generalized monoplacophoran. 

From ontogeny r — Before torsion the cup of the larval shell deepens 
with flexure of the intestinal tract and because of more rapid growth 
at the posterior margin takes the form of the beginning of an iso- 
strophic or bellerophonlike coil but with the apex or rudimentary coil 
forw^ard. It seems reasonable to suppose then that the descendants 
of our hypothetical ancestral gastropod may have passed through 
similar stages in the initial process of becoming coiled. Indeed coil- 

NO, 13 



ing could hardly have occurred in any other way. The development 
of a higher and higher shell and the initiation of coiling symmetri- 
cally in a plane are processes that lend themselves to gradual evolu- 
tionary development. If the fossil record is sufficiently complete, we 

Mouth * 

Stomach- --;^- 


- £luricie 





Glenidlyrn V-r 





Figure s 

Schematic restoration of a generalized scenellid treated as though it were trans- 
parent. In making the restoration there were employed the concepts of mi- 
torted bilateral symmetry suggested by the muscle scars of Archaephiala* 
a^ Left side view ; b^ from above. Except for the muscle scars, note the 
resemblance to the hypothetical primitive mollusk (fig. 4), The latter was 
suggested by Pelseneer without reference to scenellids which he regarded as 
anisopleurans (i .e., Docoghssa), 

should expect to find among the earliest gastropods forms with 
complete bilateral symmetry and a low, cup-shaped shell with an apex 
somewhat in front of the center j others that maintain the bilateral 
symmetry with a higher, narrower shell and the apex partly coiled 
forward, and still others with a complete coil, all steps necessary to a 
gradual evolution. As will be shown in the following parts of this 


paper, that is precisely what we do seem to find in the earliest fos* 
siliferous rocks. 

The next ontogenetic step, the sudden torsional twisting, is spec- 
tacular and of the highest significance. Since torsion is not a phe- 
nomenon that lends itself to gradual step-by-step development it is 
highly probable that it occurred just as suddenly phylogenetically as 
it does today ontogenetically; It is possible, if not probable, that 
torsion originated as the result of a genetic mutation having its 
phenotypic expression effective at the veliger stage of the ontogeny 
(Garstang, 1929, p, 89). This is the view that was set forth by W. 
Garstang and that has radically altered the thinking of many students 
of the Gastropoda. If Garstang's view is true, an isopleuran parent 
may have produced anisopleuran offspring. What could only be 
regarded as a monstrosity if it had gone no farther w^as so success- 
ful that the strain that carried the genes as a part of its heritage 
prevailed in competition and eventually brought into being an entire 
new subclass* (Class in accordance with the usual classification,) 

If torsion did arise suddenly in some such manner as Garstang postu- 
lated, then the adults of the first torted stock should have resembled 
their parent in every respect except that they had undergone torsion 
as larvae. They would have retained all their paired organs sym- 
metrically developed and their shells would have retained their sym- 
metry but with the apex or coil now in a posterior position. They 
would have retained other peculiarities of the parent stock such as 
the basic plan of ornamentation- The anus and pallial complex, how- 
ever, would be above the head and directed forward because of tor- 
sion, and since the immediate parent with a posteriorly directed anus 
and pallial complex had and needed no special provision for clearing 
the pallial cavity of waste products, the newly torted offspring would 
be like the parent in this respect. That is to say, it would have no 
anal emargination in the lip of the shell. In the Early and Middle 
Cambrian are shells that seem to meet these specifications. 

Since the new^ly acquired orientation results in a position of the 
anus and pallial complex that would seem to make it difficult to avoid 
fouling the ctenidia with waste products, we might expect that muta- 
tions providing a mechanism for ready disposal of the faeces and 
urine without fouling w^ould have survival value. Hence it is not 
surprising to find in Upper Cambrian rocks the first bellerophonts 
with an anal emargination. It is then present in three bellerophont 
families. The forms without this sanitary provision disappear shortly 


The significant embryological studies of neontologists were made on 
asymmetrical anisoplenran gastropods, the asymmetrical development 
following closely on torsion. Consequently we should not be sur- 
prised to find that lateral asymmetry appeared in the paleontological 
record soon after the establishment of a line of isostrophic gastropods 
(bellerophonts) with only torsional asymmetry. This expectation is 
realized in the appearance of the first known pleurotomarians in late 
Upper Cambrian rocks. 


Summarizing our inferences from neontological data we arrive at 
the following hypotheses which may be tested against paleontological 
data. The first is that the Polyplacophora and the pleurotomarians 
were derived from a common ancestor with complete bilateral sym- 
metry. We infer also that the Polyplacophora have evolved from that 
common ancestor through the segmentation of the shell but retention 
of bilateral symmetry. We may further Infer that the pleurotomarians 
have evolved first through the introduction of torsional asymmetry 
by a single mutation phenotypically effective at the veliger stage of 
ontogeny (bellerophonts) and later through the initiation of the 
lateral asymmetry that characterizes all the Anisopleura other than 
the bellerophonts. (See fig. 6.) Lateral asymmetry is carried pro- 
gressively much farther in more advanced groups. We may still 
further infer something of the probable characters of the isopleuran 
common ancestor of the Polyplacophora and the pleurotomarians and 
of the intermediate stages between the pleurotomarians and that 


The immediate predecessor of the pleurotomarians should have had 
all the characteristics of that group except lateral asymmetry. It 
should have had torsional asymmetry but lateral symmetry ; it should 
have been coiled tightly or loosely or with a curved apex pointing to 
the rear in a plane, with each side the mirror image of the other 
(isostrophic). It should have had an anal emargination when fully 
established but not when it first suddenly came into being. It, in turn, 
should have had an immediate predecessor with a high, conical shell 
with curved apex as the first step toward isostrophic coiling. The 
high conical shell would of course have been deep and would have so 
crowded the multiple paired pedal muscles that there would have been 
room for only a few, perhaps only a single pair. If torsion had taken 
place the apex of the shell would have pointed backward, if not it 
would have pointed forward. The last-mentioned stage should have 
had as a predecessor an tmtorted mollusk with complete bilateral 



symmetry and a low, shallow conical shell with little or no flexure of 
the viscera and, of course, no torsion, 

Sincej by hypothesis^ this most remote stage was ancestral to the 
Polyplacophora as well as to the Anisopleura it might conceivably 
display characters basic to the transverse segmentation of the shell 
into separate plates, characters such as multiple transversely paired 



Figure 6 

Oj Schematic drawing of a primitive pleurotomarian seen from above (modified 
from Naef, iQii). The primitive anal emargination is shown as a 
U-shaped sinus. The anal tube leading from the stomach is shown as 
passing through the pericardium and terminating close to the emargina- 
tion and between a pair of ctenidia, 

&, Similar drawing of a primitive sinuitid bellerophont (also modified from 
Naef, 1911), It is thought to have been very like the primitive pleuro- 
tomarians but with complete lateral symmetry. Both have undergone tor- 
sion and the pallial complex is anterior instead of posterior, 

pedal or shell muscles. It might have possessed in a nidimentary form 
traces of the tubules that carry the aesthetes in the tegumentum of 
modern Polyplacophora. 


In order to dispel certain misconceptions widely prevalent in neon- 
tological circles and, alas, occasionally met with in paleontological 
circles, a few w^ords in general terms about the gastropods of the 
Cambrian period may be helpfuL These misconceptions arose largely 
through the efforts of paleontologists of an earlier day to place spe- 
cies, inadequately understood because of poor preservation or some 
other cause, in established genera^ often in order to avoid erecting 


new genera for them. Since much of the evidence is not explicitly 
in the literature and cannot be introduced here without making this 
paper too diffuse, I am forced to speak dogmatically on some points. 

There are no Platyceratidae known from rocks older than Middle 
Ordovician, and no Capulidae in the Paleozoic, That there are both, 
is a common error, Pleurotomarians do not occur throughout the 
Cambrian section so far as is known, and only the Late Cambrian 
bellerophonts are readily recognized as such. There are only four 
pleurotomarian genera known from Cambrian rocks, {Plenrotomaria 
Sowerby, 1821, is not among them and indeed did not appear in the 
Paleozoic.) These four are Sinuopea Ulrich, igii (pL 2, fig, i), 
SchtBOpea Butts, 1926 (pL 2, fig, 2) ( =Rhachopea Ulrich and Bridge, 
1931^ and Roubidouxia Butts, 1926), and Dirhachopea and Taenia- 
spira, both of Ulrich and Bridge, 1931, The anal emargination is a 
deep, rounded sinus in Sinuopea and a deep angular sinus in Schisopea 
and Dirhachopea, perhaps culminating in a short notchlike slit in the 
latter. In Taeniospira there is a moderately deep slit and a typical slit 
band. All four genera are known from beds no older than the latest 
Cambrian Trempealeauan stage. 

Six typical and unquestionable bellerophont genera are now^ known 
from the Cambrian and will the neontologist please note that Bellero- 
phon Mont fort, 1808, is not among them. These are Ozvenella Ulrich 
and Scofield, 1897, and Claudia, Anconochihis, Sinuella (pi. i, fig. 
10), Strepsodisciis (pL i, fig. 8), and Chalaro strep sis (pL i, fig. 12) 
(all of Knight, 1947 and 1948). The first four have rounded sinuses 
as anal emarginations, the fifth a deep V-shaped sinus, and the last 
a deep sHt, All these are of late Cambrian age. The earliest is 
Strepsodisciis of the early late Cambrian Dresbachtan stage, and 
three of them, Strepsodisciis, Sinuella^ and Anconochilus, occur earlier 
than any known pleurotomarian genera. Also there are two isotro- 
phically coiled genera, Coreospira Saito, 1936 (pi, i, fig, 7), and 
Cycloholcus Knight, 1947, both referred to the Coreospiridae. Al- 
though neither has an anal emargination, the Coreospiridae are here 
regarded as primitive bellerophonts, Coreospira first appeared close 
to the boundary between the Lower and Middle Cambrian, probably 
on the upper side. 

There is also still another genus appearing still earlier and ranging 
throughout the Cambrian that must be considered in this connection. 
It is Oelandia Westergard, 1936, which is here placed in the Coreo- 
spiridae, It w^ll be considered more in detail on a later page. 

In addition to the bellerophont genera discussed, three genera of 
macluritoid gastropods occur in the last stage of the Upper Cam- 


brian, the Trempealeauan, Scaevogyra Whitfield, 1878 (pi 2, fig. 7), 
Maiherella Walcott, 1912 (pL 2, fig, 10), and Kobayashiella Endo, 
1937- All other known Cambrian gastropods are referable to iso- 
pleuran monoplacophoran genera. Of these Helcionella (ph i, fig, 2), 
and Scenella Billings, 1872 (pL i, fig, i), both put in their appearance 
along with Oelandia (pi i, fig. 5) in the Lower Cambrian and are 
thus among the earliest gastropods known. In addition to these there 
is that very puzzling, problematical group of gastropodlike shells, 
Pelagiella Matthew, 1S95, and its allies, that range throughout the 
Cambrian. These, for reasons given later in this paper, may be re- 
garded as an independent branch from some unknown gastropod 
ancestor or they may not be gastropods at all 

Although not yet described or announced in the literature chitons 
(Polyplacophora) are known from Upper Cambrian beds of the 
Trempealeauan stage- 




Continuing to proceed from the better known to the less well 

known, we will work backward from the living pleurotomarians, from 
which can be gleaned the basic anatomical details of the group, search- 
ing step by step for fossil forms that may be taken for representatives 
of the various stages in their evolution from their most primitive an- 
cestral stock. The living pleurotomarians are referred currently to 
the Pleurotomariidae, the Scissurellidae, the Haliotidae, and the 

In starting on our exploration it seems safe to assume that the 
basic organization of the most ancient pleurotomarian was essentially 
the same as that of its living representatives. Such a procedure per- 
mits us to drop rapidly down the gastropod family tree or backward in 
time something over 400,000,000 years to the late Cambrian when the 
first known pleurotomarians livedo continuing all the while along a 
branch that is easily recognized because its members show asym- 
metrical coiling and because of the anal emargination, a slit, sinus, or 
notch in the outer lip of the shell From this vantage point in the 
remote past we may examine our surroundings, particularly those a 
little more ancient. The objects of our search are forms that resemble 
the pleurotomarians very closely but are still more primitive* 

Contemporaneous with the earliest known pleurotomarians and in 


part preceding them are the bellerophonts.^ All but the most primi- 
tive are so very similar to the pleurotomarians in a number of signifi- 
cant particulars that on comparative anatomy alone they must be 
regarded as quite closely related. The shells of the bellerophonts are 
coiled typically in a close spiral but the coiling is isostrophic rather 
than helicoidal; the whorl cavity is, of course, very deep and the 
two symmetrical retractor muscles are inserted one on each side deeply 
virithin the aperture at the two ends of the columella in such a posi- 
tion that their retraction would withdraw the head and foot within 
the aperture ; there is an anal emargination, a U-shaped or V-shaped 
sinus or a slit, just as in the contemporary pleurotomarians. In fact 
the only obvious particular in which the bellerophonts differ from 
pleurotomarians is that the coiling is isostrophic and the shell is a sym- 
metrical spiral Clearly then, the bellerophont, like the pleurotomarian, 
w^as a prosobranch, but a symmetrical prosobranch* Since lateral 
symmetry is a primitive character in the mollusks this is precisely 
what one might expect in the immediate ancestor of the pleuroto- 
marians w^hich themselves retain more or less symmetrically paired 
organs. It is commonly believed by neontologists that asymmetry is 
an immediate and necessary result of torsion. No doubt the belief is 
well founded in the sense that torsion precedes asymmetry and is a 
prerequisite for it, but if the bellerophonts are prosobranchs as their 
morphology strongly suggests and if torsion is the factor that dis- 
tinguishes a bellerophont from an immediate laterally symmetrical 
isopleuran ancestor, then, as the time factor insists, it is not necessary 
to suppose that asymmetry was an immediate consequence. Of course 
torsion furnished the unstable condition that ultimately led to asym- 

Again surveying our surroundings, this time from the apparent base 
of the bellerophont stem, we meet with two more genera that have 
the characters one would expect of the very primJtive bellerophonts. 
One is Cycloholcus from the base of the Upper Cambrian Dresbachian 
stage and the other is Coreospira (pL i, fig. 7) (both referred to 
previously) from close to the boundary of the Middle and the Lower 
Cambrian, probably on the upper side of the boundary. Both of these 
forms are isostrophically coiled and thus in this respect are in accord 

9 Some views expressed by Thiele, 1935 (p. 1125), and Wenz, 1938 (pp. 58-60), 
on the probable anatomy and physiology of the bellerophonts will not, I think, 
bear close scrutiny. Since I do not wish to interrupt the present argument to 
give the reasons for my contrary views that the bellerophonts are prosobranchs 
instead of primitively orthoneurous "Amphigastropoda/' as Thiele and Wenz 
supposed, I am discussing the matter in an appendix to this paper. 


with the bellerophonts. Unhke previously recognized bellerophonts 
there is no emargination in the lip that corresponds to the anterior lip. 
This appears puzzling unless we remember that there is a feature 
we were to look for in the primitive bellerophont. 

Further exploration turns up the genus Oelandia (pL i, fig. 5), 
a genus that may be interpreted most plausibly as being closely re- 
lated to Cydoholcits and especially Coreospira (pL i, fig, 7), Oelandia 
has been associated commonly with Helcionella Grabau and Shimer, 
1909 (pi I, fig. 2), For example, Wenz in 1938 (p. 88) places it in 
the subfamily HelcionelHnae in the family Tryblidiidae, There is 
indeed a resemblance— a resemblance that appears to me to be hon- 
estly come by but still not decisive taxonomically. In Helcionella the 
apertural margins are in a flat or nearly flat plane. In Oelandia how- 
ever the margins tend to be curved and one end, the end toward 
which the apex bends, is considerably extended and often tilted up 
as though to form a trainlike hood. If one attempts to think in terms 
of soft anatomy this hood seems anomalous over the head but fits 
nicely as a hood over the posterior train of the foot, lience the ex- 
tended or up- til ted end is here regarded as posterior. If this hypothe- 
sis is accepted the apex is posterior and Oelandia may be considered 
to be a very primitive isostrophic prosobranch gastropod in the Cor- 
eospiridae, one that has not yet advanced to the stage of close coiling. 
Of course the anal emargination has not yet appeared. Helcionella 
remains in the Isopleura with the nontorted Monoplacophora, Oelandia 
is an anisopleuran that may have been derived directly from Hel- 
cionella and retains its characteristic ornamentation. Torsion may 
have first taken place between these two genera in earliest Cambrian 
or in pre-Cambrian time. This possibility will be discussed again. 

The Coreospiridae are bellerophonts in respect to the shell coiled 
or nearly coiled with lateral symmetry in a plane. In some other 
respects they resemble more closely the group that we next meet 
with, for although we have reached in Oelandia (pi i, fig, 5) close 
to the beginning of the fossil record we have not fully surveyed its 
contemporaries. There are still three kinds of gastropods or sup- 
posed gastropods represented with Oelandia in the Lower Cambrian 
rocks. One of these three, Pelagiella and its allies^ seems anomalous 
from any viewpoint and will be reserved for later discussion. The 
other two fit into our picture very nicely. Both are cup-shaped and 
show complete bilateral symmetry. Their ornamentation consists of 
transverse undulations somewhat similar to those of Coreospira (pL I, 
fig. 7) and Oelandia. One, the genus Scenella Billings, 1872 (pL I, 
fig, I ) , is cup-shaped with a conical shell and the apex tipped toward 


the narrower end. The shell of the other genus, Helcionella (pi i, 
fig. 2), is also cup-shaped, and includes species that are low and 
broad as well as others that are high and narrow. In both the apex 
points toward the narrow^er end of the aperture and in the high and 
narrow species it is almost hooked. None of these have the hoodlike 
train of Coreospira, Oelandia, and narrow bellerophonts in general. 

Although we know nothing of the internal organization of either 
Helcionella (pi. i, fig. 2) or Scenella (pi i, fig. i) by direct ob- 
servation, their external features such as shape and ornamentation 
suggest rather strongly that they belong to a family that continues 
into the Devonian, Specimens of an Ordovician genus of this 
family, Archaeophiala^^ Perner (pi i, fig, 3), preserve the muscle 
scars beautifully. The scars are strongly pigmented and for that 
reason are unusually sharp and clear, (See Knight, 1941, pi 3^ figs. 
3a-b.) These scars are 12 in number and are arranged in a ring deep 
within the margin of the shell. Two of the scars are larger than the 
others and are made up of three parts. These tripartite scars, which 
occur at one end, may be regarded as compound and perhaps as rep- 
resenting the scars of three muscles each. The other 10 scars are 
simple and probably are the scars of single muscles. These 12 (or 16) 
scars are in bilaterally symmetrical pairs. The pair of large compound 
scars lies at the end toward which the apex lies and very nearly closes 
the circle at that end. The scars of the other five pairs follow sym- 
metrically on either side until the circle is nearly closed at the other 
end. There is a line of much fainter, unpigmented scars outside of 
the principle ring. The six (or eight) pairs of pigmented scars were 
probably points of attachment for symmetrically paired muscles con- 
necting the shell to the foot. One can hardly guess what function was 
served by the muscles that made the more obscure scars outside those 
of the main circle but these shadow scars appear to be characteristic 
of the group. 

Two exceedingly important inferences are suggested by the scars 
of Archaeophiala (pi ij fig. 3). The first inference is that the soft 
anatomy was bilaterally symmetrical throughout, that is to say the 
animal had not undergone torsion. This is an inference primarily 
from the complete bilateral symmetry of the paired muscle scars, 

10 I am employing Archaeophiala rather than Tryblidmm to typify the gastro- 
pods with paired muscle scars for the reason that its shape, which is essentially 
that of Scenella and the lower, cup-shaped Helcionella, suggests that it is the 
more primitive. Although their muscle scars are virtually identical, I am plac- 
ing each in a separate family, as will be seen, since each seems to be a mem- 
ber of a different series, each with its characteristic shape. 


supported by the lack of an area between scars at either end for a 
pallial cavity. The second inference is that the end that has the large 
compound muscle scars and toward which the apex lies is anterior. 
This follows as probable from a corollary to the principal of cephali- 
zation to the effect that "heteronomoiis segmentation is an expression 
of cephalization." If one takes these two inferences together with 
the previous inference that such Cambrian genera as Scenella (pi. i^ 
fig, i) and at least the lower, cup-shaped species of Helcionelia (pL i, 
fig, 2) are organized in a similar way we have a working hypothesis 
as to the organization of these very important early forms. It seems 
quite certain that the superficial resemblances of these Cambrian cup- 
shaped forms to the living prosobranch patellaceans or capulids or 
to the equally prosobranch Paleozoic platyceratids is as surely a mat- 
ter of convergence as is the equally superficial resemblance of all of 
them to the pulmonale ancylids. 

It seems probable that a prerequisite for torsion was a reduction 
in the hypothetical six or eight paired shell muscles to a single pair< 
What better mechanism to give mutations accomplishing such a re- 
duction survival value could have been devised than the development 
of high, narrow shells, such as actually occurred in some Lower and 
Middle Cambrian species currently referred to Helcionelia (pL I, 
fig. 4)- In these the hypothetical six or eight pairs of muscles, if 
present J would be crowded together. Perhaps, owing to this crowd- 
ing, mutations that would effect the reduction of the six or eight pairs 
to a single pair through the elimination of all but one of the pairs 
would have survival value. If the suggested reduction actually took 
place the foundation was laid for torsion. All that would be required 
further is that through a genotypically small mutation the rudiments 
of one muscle of the pair (the left one) should develop in the early 
veliger larva earlier than those of the other. As has been shown by 
Crofts (1937), the retraction of such a single asymmetrical "velum 
retractor muscle" in the early veliger is what actually initiates torsion 
in Haliotis, Undoubtedly when torsion first appeared in the remote 
ancestors of Haliotis the same mechanism w^as responsible for it. 

Both Helcionelia (pL i^ figs. 2 and 4) and Scenella (pi. i, fig. i) 
appear in Lower Cambrian rocks. It seems probable that Helcionelia 
and Scenella had a common ancestor in early Cambrian or in pre- 
Cambrian time. Chuaria Walcott^ from pre-Cambrian rocks of the 
Grand Canyon region, has been suggested as the most primitive an- 
cestral gastropod but the only known specimens of the only known 
"species/' all of which I have examined, are so very poorly preserved 
that it is utterly impossible for me to recognize them as gastropods or 


anything else. The most I can say of the specimens is that they may 
be organic in origin. 

In descending the family tree we have passed from the earliest 
forms that can be assigned to the Anisopleura with assurance, the 
bellerophont cyrtolitids and sinuitids, such as Strepsodiscus (pL i, 
fig. 8) and SinueUa (pL i, fig, lo) of the lower and middle Upper 
Cambrian, through the probably anisopleuran Coreospiridae, to the 
isopleuran Helcionella (pi. i^ fig. 2) and Scenella (pL i, fig. i) of the 
Lower and Middle Cambrian. In doing so we have passed along two 
exclusively Cambrian limbs, the Coreospiridae and the Helcionellidae, 
The Coreospiridae resemble the bellerophonts externally except that 
there is no feature that can be assigned the function of the bellero- 
phont anal emargination. The Helcionellidae resemble the Coreo- 
spiridae except that the direction tow^ard which the apex bends is 
interpreted as anterior. As stated above, there are with the genus 
Helcionella (pL i, figs. 2 and 4), as currently understood, species that 
have a high shell with a strongly curved apex and others^ more simi- 
lar to the type species, with a low shell with the apex so short and 
blunt that in some specimens it is almost an overstatement to say that 
it is curved at all. These appear to make a continuous series. Our 
hypothesis requires that torsion was initiated somewhere between 
the untorted helcionellids and torted bellerophonts. The evidence for 
one point in the chain as against another is not very compelling. I 
have placed the dividing line between Helcionella (pL i, fig, 2) and 
Oelandia (pi i, fig, 5), placing the former in the Isopleura with the 
Monoplacophora and the latter in the Anisopleura with the bellero- 
phonts. If anyone prefers to class the Coreospiridae with the Mono- 
placophora or Helcionella with the bellerophonts, I cannot quarrel 
too vigorously with the preference. There is insufficient evidence. As 
the muscle scars, which might give more objective evidence, are un- 
known in Helcionella and in both Oelandia (pL i, fig, 6) and Coreo- 
spira (pL I, fig, 7) we are left with little but interpretations from 
w^eak morphological data as basis for a decision, however tentative. 
What little objective evidence there is lies in the similarity of the 
ornamentation in the Helcionellidae and the Coreospiridae and in 
differences in the apertural margins. This suggests that both of them 
are allied to each other and to the Scenellidae where the ornamental 
tion follows a similar pattern, but that for some reason, assimied to be 
torsion, the apertural margins are different. At whatever point tor- 
sion was introduced, our hypothesis requires that it was in the more 
or less advanced descendants of Scenella and Helcionella that con- 
ceivably retained a similar type of ornamentation. 



Just as the neontologists have employed restorations of the hypo- 
thetical primitive mollusk with fruitful results, so the paleontologist 
with even more actual data, the fossil shells, may employ them also. 
Not only does the paleontologist have fossil shells that tend to sup- 

Ventrfcle- - 
Auricle - 


Clenidlum — 




- -WCtenidium 




Figure 7 

a. Schematic restoration of Coreospira as a monoplacophoran isopleuran. 

fc, Schematic restoration of Coreospira as an isostrophic anisopleuran, a hel- 
ler ophont without an anal emargination. The latter seems a much more 
plausible restoration than the former* Of course, neither restoration may 
approximate the truth, but in that case Coreospira would probably not 
have been a gastropod. It is understood that such organs as ctenidia, 
auricles, etc*, are paired in both restorations. The probable retractor 
muscles are not shown. 

port the scientific speculations of the neontologist but he has others 
to which he may attempt to fit the soft parts of a generalized gastro- 
pod and form judgments from the plausibility of the results as to 
what the animal as a whole may have been like* Some of these hypo- 
thetically restored gastropods tend to fill gaps between the untorted 


monoplacophoran and the torted bellerophont which in turn connects 
closely with the pleurotomarians, 

For example, text figure 7 shows two restorations based on the 
known shells of Coreospira (pi, 1, fig, 7), Figure 7, a, shows the shell 
and hypothetical soft parts restored as an isopleuran monoplacophoran. 
Figure 7, b, shows the same restored as an anisopleuran bellerophont. 
Obviously the second yields a plausible picture of the probable re- 
lationship of shell and soft parts. It looks comfortable. The mono- 
placophoran restoration is too fantastic for even tentative acceptance- 
Even though one should restore the soft parts to display more primi- 
tive isopleuran features, a row of muscles, a very shallow posterior 

^ stomach 


1..-Z4- Ctenidium 


ii-X' Mouth 

Figure 8 

Oclandia restored as a bellerophont even more primitive than Coreospira, It 
presents a harmonious and plausible picture* A restoration as a mono- 
placophoran is quite as unacceptable as is the same restoration of Coreospira. 
As I have pointed out previously the trainlike hood over the posterior part 
of the foot is a critical feature — a feature that is shared with several 
bellerophont genera with narrow shells. 

pallial cavity, etc*, we still would have the coiled shell suspended 
above the head in a most unacceptable fashion, as well as a narrow, 
coiled visceral mass entirely incongruous on a monoplacophoran. 
Surely it is difficult to accept Coreospira as other than a primitive 

Figure 8 shows a restoration of Oelandia (pi. i, fig. 5), believed 
to be a bellerophont even closer to the Monoplacophora than Coreo- 
spira (pL I, fig. 7), 


The procedure of working backward may be likened to selecting 
one terminal twig of a tree from among very many, a twig on a 


branch that by preliminary inspection took its origin far down on the 
trunkj and then following that branch still farther down until one 
is led by the process to what appear to be the roots. But our meta- 
phorical tree, from preliminary inspection in very poor light (for let 
Its assume that we are feeling our way in the dark of the moon), 
seems to have more than one main branch. One of these which we 
will call the Polyplacophora, appears near the roots to lie close 
to the branch that we have been tracing backward w^ith apparent 
success. Let us examine it further. 

The chitons or Polyplacophora, far from abundant today^ have al- 
ways been rare in the fossil record. Nevertheless they are reported in 
the literature as distributed throughout geological time from rocks 
as early as Lower Ordovician and in the collections of the United 
States Geological Survey housed in the United States National Mu- 
seum are specimens of unquestionable poly placopho ran plates from 
the late Upper Cambrian Eminence dolomite of Missouri, These 
specimens, belonging to species and perhaps genera yet undescribed, 
are nevertheless typically polyplacophoran in every detail including 
the peculiar surface sculpture, common to all chitons, possibly to be 
associated with the remarkable shell eyeSj, or aesthetes, developed in 
this group. 

Thus our leap backward in time along the polyplacophoran limb of 
our metaphorical tree carries us almost exactly as far as our leap 
along the pleurotomarian limb, to latest Cambrian time. However, we 
find no obvious intermediate connections with any monoplacophoran. 
Our only clue appears to be offered by the paired multiple dorsal 
muscle scars of Archaeophiala (pi. I, fig. 3) attributable by analogy 
to Helcionella (pk i, fig< 2) and Scenella (pi. i, fig, i), possibly rein- 
forced by what appear to be tubules in Tryblidium very similar to those 
w^hich carry the nerves for the aesthetes in chitons. In Archaeophiala 
(and in Tryblidium) the number of pairs is six ^^ but the pair of 
large scars at the end regarded as anterior are compound and made 
up of three smaller elements so that the basic number of pairs might 
be regarded as eight. One might infer that the eight-segmented shell 
of the polyplacophoran was merely the single shell of the mono- 
placophoran separated into eight segments to correspond with the 
eight pairs of shell muscles, 

^^ It may be significant that the embryos of living polyplacophorans first 
develop six shell plates. The other two, the terminal plates, are added at a later 
stage (Garstang, 1529, p. 78), 



For recapitulation it may be well to reverse our course and sum- 
marize our results by ascending the pleurotomarian branch of the 
family tree beginning with the Monoplacophora, We will still hold 
to this one line, lest we go entirely astray, and we will arrive at the 
present-day level along a limb with nothing more advanced than the 
highly specialized relics of the once great pleurotomarian stock. 

Throughout rocks of Cambrian age we find what appear to be 
primitive gastropods with low; cuplike shells. The apex is subcentral 
or anterior and there is no posterior train. All have rather coarse 
transverse plicae or costae and finer ornamentation as well They 
are believed to have six (or eight?) symmetrical pairs of adductor 
muscles and not to have undergone torsion. Typical of these early 
Cambrian genera are Scenella (pL i, fig- i) and those species of 
Helcionella (pL i, fig. 2) that have the low cuplike form of the 
genotype. Probably these or similar forms were in existence in late 
pre-Cambrian time. 

Concurrent with the more typical species of Helcionella (pb i, 
fig. 2) are other species, that should probably be referred to another 
as yet unnamed genus, which have very high, narrow shells (pL i, 
fig. 4), It is possible that in these forms the adductor muscles were 
so crowded that their number was reduced to a single pair, seemingly 
a prerequisite for the initiation of torsion. Likewise in the early two- 
thirds of Cambrian time are found species of Oelandia (pL i, fig, 6), 
much like Helcionella externally but with an extended or up-tilted 
margin under the apex that has the same shape as the posterior train 
found in narrow bellerophonts. Accepting it as homologous, we then 
must accept Oelandia (pi i, fig- 5) as having undergone torsion but 
in most other respects to have retained at least some of the external 
features of Helcionella, It is possible that it was the first bellerophont 
and first prosobranch. Its apex is posterior but still not truly coiled. 
However, close coiling is found in Coreospira (pi i, fig. 7) partly 
contemporaneous with all these but appearing first a little later and 
still without the anal emargination. 

In late Cambrian time we find a number of bellerophonts each pro- 
vided with an anal emargination: StrepsodisciiS (pi i, fig, 8) and 
Cloudia in the Cyrtolitidae ; Sinuella (pi i, fig. 10), Owenella, and 
Anconochihis in the Sinuitidae; and Chalaro strep sis (pi i, fig- 12) 
in the Bellerophontidae, With this beginning the bellerophonts de- 
ploy throughout Paleozoic time and have their last representatives 
in the Triassic, 


In the latest Cambrian the first pleurotomarians put in their appear- 
ance, mostly primitive pleurotomariaiis with either rounded or angular 
sinuses, Sinuopea (pL 2, fig. i), Schizopea (pL 2^ fig, 2), and Di- 
rhachopea. The anal emargination in Taeniospira is a true slit. The 
anal emargination became a true slit in Early Ordovician time 
in a number of genera, some as yet undescribed. During the re- 
mainder of the Paleozoic the pleurotoniarians proliferate greatly 
and seemingly gave direct rise independently to a number of non- 
pleurotomarian aspidobranch stocks and through these to most if not 
all of the more advanced gastropods. They continue in declining num- 
bers through the Mesozoic and Cenozoic and survive today in greatly 
reduced numbers as relic families adapted to special environments. 

In late Cambrian rocks in beds almost contemporaneous with those 
containing the earliest known pleurotoniarians^ the first known poly- 
placophorans appear, typical chitons in all respects. These also con- 
tinue to the present day but always as relatively few forms mostly 
adapted to rock clinging. They, like the anisopleuran branch, seem 
to have been derived from primitive, untorted monoplacophorans but 
through an entirely different set of modifications. The primitive iso- 
pleuran condition continued, for in the polyplacophorans there w^as no 
torsion, but the primitive single cuplike shell is replaced by eight 
transverse plates. Perhaps these eight plates represent the primitive 
shell which may have become divided transversely in accordance 
with the possibly eight pairs of shell muscles, 


In our climb down two branches of the gastropod family tree, arriv- 
ing along both at the same main stem, we have followed what appears 
to be a logical and straight course, paying no attention to other 
nearby branches. But there are other nearby branches not too far 
above the roots and it would be improper to leave them out of con- 
sideration altogether, especially as the light is very poor, 


First there is the branch that we will call the Patellacea. It is well 
represented in our living faunas and goes far back into geological 
time. The patellaceans include simple, cuplike shells that show ex- 
ternally full bilateral symmetry and resemble very closely those that 
we are here regarding as monoplacophoran isopleurans. However, 
the anatomy and ontogeny of living representatives show unequivo- 
cably that the symmetry of the shell is superficial and secondarily 


derived. They are classified by neontologists as Archaeogastropoda in 
the Prosobranchia. Although primitive in many respects, they show 
in their soft parts and in their ontogeny both torsional and lateral 
asymmetry. Can it be that in following our branch backward in time 
we have become confused in the darkness and, instead of passing 
from the earliest bellerophonts to monoplacophorans, we have stepped 
across onto another superficially very similar but different branch? 
Can it be then that what we are calling monoplacophoran isopleurans 
are in truth nothing more than very ancient patellaceans ? Except 
for Wenz, most previous authors have so regarded them. However, 
I think not, for there are characters in both groups, very obscure ones 
to be sure, that seem to indicate the contrary, 

The significant clues have to do with the scars of the shell muscles 
in each group. In the patellaceans the muscle scars form a continuous 
horseshoelike crescent, open anteriorly, for the shell muscle does not 
intrude upon the region occupied by the anterior pallial cavity. The 
shell muscle is composed of closely applied bundles of muscle fibers 
and in some species this is reflected in the scar by knots, so to speak, 
in the continuous scar that suggest the discrete scars of the typical 
monoplacophoran s. But these knots in the patellacean scar are not 
symmetrically paired while the discrete scars of the monoplacophorans 
are. They reflect the basic asymmetry of the patellaceans. Likewise 
the anterior opening of the patellacean scar seems to reflect the torsion 
of the primitively posterior pallial complex and cavity to an anterior 
position above the head. Although there is in the patellaceans a very 
thin scarlike line connecting the open ends of the horseshoe, it is 
apparently not the scar of the pedal muscle but merely the line of 
attachment of the mantle to the shell, analogous to the pallial line 
of the pelecypods. The monoplacophorans are here conceived to have 
included also forms with a continuous muscle scar, such as Archina- 
cella Ulrich and Scofield, 1897, as well as those with discrete paired 
scars, but in both types the scars have elements that close or nearly 
close the circlet anteriorly and these elements seem to be continua- 
tions of the scars themselves. This suggests that these forms, like the 
Polyplacophora, do not have an anterior pallial cavity and supports 
our inference that the Monoplacophora have not undergone torsion. 
That the scars are narrower anteriorly may be accounted for if we 
imagine that the muscles attached at this part are extensions from 
the pedal muscles at each side arching over the head. 

Although Wenz did not recognize them as such, it is my opinion 
that the late Paleozoic genera, Metopfotna Phillips, 1836, and Lepe- 
topsis Whitfield, 1882, are not monoplacophorans but are referable to 


the Patellacea. Both have continuous, horseshoe-shaped muscle scars^ 
completely open at the anterior end. I would also assign Palaeo- 
scurria Perner, 1903, to the Patellacea. Per net described and figured 
for this genus an open horseshoe of almost discrete muscle scars but 
I have examined the types of his genotype species and can find no 
objective evidence for the existence of such a feature (Knight, 1941, 
p. 231). Nevertheless, there is no direct evidence for any other sort 
of scar^ possibly because the matrix is too coarse to record such deli* 
cate features. However, the shape of the shell is so similar to that 
of Le pet apsis that I shall provisionally associate the two. The fossil 
record of the Patellacea is then continuous from at least Mississippian 
and perhaps from Silurian time to the present. I know of no Patel- 
lacea from rocks earlier than Silurian, nor do I know any forms 
transitional from pleurotomarian to patellacean unless the very im- 
perfectly known Halophiala Koken, 1925, from Ordovician rocks 
may be so regarded. 


Beginning in the early Trempealeauan stage of the Upper Cam- 
brian and ranging into the Middle Devonian are a series of genera 
that give the appearance, at least, of being coiled sinistrally. These 
genera are here united taxonomically not only by the apparent sinis- 
tral coiling, but by another feature as well. This feature, a difficult 
one to describe, consists in most of these genera of a peculiarity of 
the region surrounding the umbilicus or that part of the shell usually 
called the base whereby the ^'basal" part of the whorl profile is rather 
sharply arched, most conspicuously so where there is an open um- 
bilicus. This sharp arching of the supposed basal part of the whorl 
resembles a notch keel with an internal channel. In many forms it 
is clearly the locus of a sinus in the lip. The following 17 genera, 
most of them commonly regarded as sinistral, make up the group I 
have in mind ■ 

From the Upper Cambrian racks 

Kobayashiella Endo, 1937. 

Matherella Walcott, 191 2 {ph 2, fig. 10), 

Scaevogyra Whitfield, 1878 (pi. 2, fig. 7). 

From Ordovician rocks 

Aniispira Perner, 1903. 

BarneseUa Bridge and Cloud, 1947 (p. S4S). 

Clisospira Billings, 1865, 

HelicoHs Koken, 1925, 


Laeogyra Perner, 1903. 

Lecanospira Ulrich, in Butts, 1926 (ph 2, fig. 8). 

Lesucurilla Koken, 1898. 

Maclurites LeSueur, 1818 (pi. 2, fig. 12) (=zMaclurina Ulrich and 

Scofield, 1897). 
Macluriiella Kirk, 1927. 
Mather ellina Kobayashi, (1933) 1937, 
Mimas pira Koken, 1925* 

Palliseria Wilson, 1924 (pi, 2, fig. 11) {::=Miirospira Kirk, 1930) < 
Vers is pira Perner, 1903. 

From Silurian rocks 

Onychochiius Lindstronij 1884 (pk 2, fig. g) (= Palaeopupa Foerste, 

From Devonian rocks 

Sinistracirsa Cossman, 1908 (^Dormldia Perner, 1903, preoccupied, 

and Boycotfia Tomlin, 1931). 
Omphalocirrus Ryckholt, i860 (i^ Coelocentrus Zittel, 1882, Poly- 

enaulus Ethridge, 1917, and Arctamphahis Tolmachoff, 1926), 

Thus I have grouped together (with one or two superficially dex- 
tral genera) all the Paleozoic genera commonly regarded as sinistral 
except Antitrochiis Whidborne, 1891, which I refer tentatively to 
the Trochonematacea, Agnesia Koninck, 1883, and Hesperiella Holz- 
apfel, 1889, both pleurotoniarians and possibly congruent, and Cam- 
bodgia Mansuy, 1914, a pseudomellaniid. Other typically dextral 
genera are known to have a few sinistral species, as w^ell. 

Up to this point I have spoken of the gastropods of the group we 
are considering as '^apparently" sinistral, that is to say, when the 
shell is oriented in the arbitrarily conventional position ^^ with the 
spire upward (or the umbilicus downward), the aperture is below 
and to the left rather than to the right as in the vast majority of 
gastropods. In the truly sinistral gastropod all organs of the body 
are reversed in position from that of the dextral gastropod beginning 
ontogenetically with the early cleavages of the egg. The reversal 
appears to be the result of a mutation that may occur in some indi- 
viduals of normally dextral species, or that has become fixed in the 
heritage of some species in genera that are otherwise dextral, or of 
a few entire families. 

Sinistrality is well known among living gastropods but relatively 
it is very rare. Likewise it is known among fossil Gastropoda. Of 

12 I employ the illogical conventional orientation preferred by English, German, 
and American authors. 


the genera dealt with above Antitrochns and Cambodgia are probably 
sinistral- The pleurotoniarians Agnesia and Hesperiella present a 
still different picture tliat I hope to discuss at another time. But in 
all cases sinistrality is a deviation from the basic plan and seemingly 
occurs only as the result of mutations that may or may not become 
fixed in the heritage of a group^ It occurs sporadically in various 
only remotely related groups and is probably of no selective value, 
positive or negative, to its possessor. The rarity of sinistral gastro- 
pods is related to the primitive torsion and asymmetry of the Aniso- 
pleura. Presumably it was of such a nature as to produce dextral 
forms, and deviations from the plan require a relatively rare muta- 
tion in which all parts of the organism at all stages were reversed. 
Therefore the occurrence of a relatively large number of apparently 
sinistral forms classifiable into a relatively large number of genera 
very early in the history of the class is startling and affords grounds 
for suspicion that these forms are not truly sinistral. 

Among living gastropods there is another phenomenon very much 
rarer than sinistrality which gives rise to a shell that has the appear- 
ance of being sinistral but the organs of the anatomy are not reversed 
from the position in dextral Gastropoda. The entire animal, including 
both soft parts and shell, is actually dextral in this case, and the 
shell is ultradextral or hyperstrophic, not sinistral. In other words, 
the normal spire has sunk inward, as it were, and may even be coiled 
in such a way as to produce an umbilicus. The normal base may be 
flat or protrude to resemble a spire in every respect. Hence a shell 
is produced that appears to be sinistral although it is actually dex- 
tral The *'spire'' of such a shell is homologous with the base of what 
may be called a normal dextral shell and its "base" is homologous 
with the spire* (See fig. 2.) 

Hyperstrophy is exceedingly rare among living gastropods^ occur- 
ring most frequently as a specialization only in the embryonic nucleus 
of some opisthobranch gastropods and in the adult stage of a few end 
members of various highly specialized groups such as pteropods, 
Ampullariidae, and pulmonates. If it were to occur in adults of spe- 
cies with unknown soft parts it would be difficult to distinguish from 
sinistrality except on collateral evidence. One line of collateral evi- 
dence is that supplied by the peculiar angulation on the "base" of 
these early Paleozoic shells. If we regard these shells as hyper- 
strophic, the angulation is no longer anomalous. It becomes the trace 
of the dorsal anal emargination. There is another line of collateral 
evidence that is exceedingly pertinent to at least one of the Paleozoic 
genera that is included in the group we are discussing and it seems 


very strong evidence indeed. This evidence is furnished by the oper- 
culum virhich is preserved in this genus because it is calcified- 

The gastropod operculum is basically corneous ( conchy olin) but in 
some groups the corneous operculum is partially or wholly calcified 
and in some forms this makes the operculum very massive. Only 
where it is calcified is the operculum of fossil forms preserved. The 
embryonic operculum is a minute disk that grows by incremental addi- 
tions to a margin or margins. Where the increments are added mark- 
edly to one side of the margin as against the other sides, growth may 
be in a spiral and such a spiral operculum, as seen on the external 
face of the operculum in dextral gastropods, always grows from the 
nucleus in a counterclockwise direction. In sinistral gastropods it is 
clockwise. Now, in one of the genera of Paleozoic gastropods of the 
group we are considering, the operculum was thick and calcified, 
consequently it is not only frequently preserved but in some speci- 
mens it has been found in place in the aperture. The genus is Mac- 
lurites LeSueur, 1818 (pi. 2, fig. 12). In Maclurites the operculum 
is in the form of an open spiral and the direction of coiling, as seen 
on its outer face, is counterclockwise. Hence, as pointed out by S, P. 
Woodward as early as 1854 (p. 202), the shell of Maclurites is not 
sinistral, as has often been supposed, but dextral and hyperstrophic. 
We do not know the operculum in any other of the genera included 
in the group under consideration. Nevertheless, as I have endeavored 
to show, the group w^e are considering appears to be a natural unit 
and we may therefore with reasonable assurance attribute to the 
other genera the char?xter of hyperstrophy that the angulation on 
the ''base'' of the whorls suggests and that the operculum of Mac- 
lurites seems to confirm. 

Assuming that we have solved the problem of the coiling in 
the group under consideration, namely that it is hyperstrophic dex- 
tral rather than sinistral, we are faced in consequence with an even 
more difficult problem. What is the meaning in terms of soft anat- 
omy and of phylogeny of a rather large group of dextral hyper- 
s tropic forms introduced so very early in the history of the Gas- 
tropoda? Since this group seemingly became extinct before the 
close of Devonian time, it left no recognized descendants among liv- 
ing gastropods that might throw some light on its organization. As 
suggested above, hyperstrophy is very rare among living gastropods 
and occurs only as a secondary acquisition in groups far removed 
from any possible connection with our early Paleozoic group. We 
can only surmise what the anatomy of the soft parts of the Paleozoic 
forms might have been. The shells are coiled and coiled asymmet- 


rically. These facts suggest torsion and possible asymmetry in the 
primitively paired organs. But, as the shells are hyperstrophic and 
appear in the fossil record shortly before the first known pleuroto- 
marians, the asymmetry possibly may be very different from that of 
the main line of gastropod descent. Figure 9 shows hypothetical 
restorations of the hyperstrophic genus PaUiseria (pi. 2, fig. 11) a 
close relative of Maclurites (pi. 2, fig. 12). Accepting the notch keel 
surrounding the umbilicus as the locus of the anus we find very little 
room for a right ctenidium and tentatively assume that this and its 
associated organs had been lost. The operculum of Maclurites is not 
only that of a dextral shell but it shows a startling resemblance to 
that of the recent Nerita in that there are points of attachment for a 
pair of retractor muscles. A single pair of retractor muscles is a 
primitive feature shared with the bellerophonts, the more primitive 
pleiirotomarians and the neritaceans. The line of speculation that 
seems most plausible to me is that this group branched off from the 
early bellerophonts at some such stage as is represented by Strepso- 
discus (pi. I, fig. 8), a bellerophont that commonly shows some asym- 
metry in a sinistral or hyperstrophic sense. This would accord with 
chronogenesis^ for Strepsodisats precedes in the fossil record Scacvo- 
gyra (pi. 2, fig, 7), the earliest hyperstrophic gastropod, and both 
precede the earliest pleurotomarian. If this is true the somewhat 
angular ''base^' of Scaevogyra^s whorl is homologous with the angular 
dorsum of Strepsodiscus. It is further supposed that asymmetry 
arose in the group under consideration as an early genetic response 
to the mechanical difficulties of isostrophic coiling as in the main line 
of gastropod descent that began with the pleurotomarians, but inde- 
pendently and probably in a somewhat different way, and that this, in 
turn, resulted in the asymmetrical, hyperstrophic shell, 


Doubts have been expressed that the members of this group are 
actually gastropods in spite of the very close resemblance of their 
shells to those of gastropods. Thus Wenz wrote in 1938 (p. 95) of 
Pelagiella: "Systematische Stellung Fraglich; vermutlich uberhaupt 
nicht zu den Gastropoden gehorig/' The late Dr. E, O, Ulrich is 
said to have held the opinion that they were ''pteropods;' that is to 
say^ allied to the hyolithids,^^ The hyolithids are no longer regarded 
as true pteropods or even as gastropods for that matter. The true 
pteropods are highly specialized opisthobranch gastropods of Tertiary 

1^ Oral communication from Dr. Josiah Bridge, 

NO. 13 





Figure g 

Three views of a restoration of Palleseria longivdli (Kirk), about Xi 

a. View of the left side. Note the direction of coiling and the protruding, 
spirelike base. 

tj Anterior view, 

€, View of the right side. Note here and on the anterior view (t) the 
umbilicus occupying the side where a spire would be in a dextral ortho- 
strophic gastropod. Note especially that the ridge surrounding the 
umbilcus is the locus of a notch. This notch, an anomalous feature if 
the shell is regarded as sinistral, is believed to be the anal emargination. 
If this is correct there is very little space for the primitive right ctenidium 
and probably it has been lost. 


and Recent times, in many respects the farthest removed from the 
primitive stock of any of the class, 

Matthew, in erecting the genus in 1895, seemed to regard Pelagiella 
as a heteropod. The heteropods are again a highly specialized group 
of late Cretaceous to Recent times, although being prosobranch, not 
quite so far removed morphologically from the primitive stock as 
the opisthobranch pteropods. These quite unacceptable assignments 
serve to accentuate the difficulties in finding a place for these forms 
in the Gastropoda, My own difficulties derive from the fact that if 
Pelagiella and its allies are gastropods, the hypotheses I have been 
setting up cannot include them, except peripherally. Otherwise these 
hypotheses must be abandoned or extensively modified in such ways 
that they would meet with greater difficulties in other directions than 
the difficulty presented in removing this obviously questionable group 
to a peripheral position or eliminating it from the Gastropoda alto- 

Pelagiella and its allies occur in company with the first recorded 
monoplacophoran gastropods in the rocks of early Cambrian time and 
are not only coiled but asymmetrically coiled. Thus, if they are 
gastropods, they appear superficially to have advanced well beyond 
the isopleuran monoplacophoran stage, the anisopleuran isostrophic 
stage, and even beyond the pleurotomarian stage of the main line of 
gastropod evolution* Therefore, since they are contemporaneous 
with the earliest known monoplacophorans, and the earliest and most 
primitive known representatives of one of the two hypothetically 
more primitive stages and precede the others by a considerable in- 
terval of geological time, the difficulties are obvious. 

I do not wish to consume space in laboring the problem at too great 
length for, with our present lack of knowledge of their anatomy and 
even of their conchology, it is not soluble* I am not prepared to 
abandon the hypotheses as to the derivation of the main lines of 
gastropod descent until other hypotheses are presented that better 
explain the observed facts. Although it is recognized that the fossil 
record is imperfect, I am not prepared to assume that it is so very 
faulty that its bearing on the broader aspects of chronogenesis is to 
be set aside. There are various other possibilities that might be called 
upon to account for Pelagiella and its allies as gastropods. For ex- 
ample, can it be that they represent a branch from the monoplaco- 
phoran stock that acquired torsion and asymmetry independently in 
pre-Cambrian time, perhaps bypassing an isostrophic stage? There 
is little evidence one way or another but I think it extremely unlikely. 
Or can it be that they are monoplacophorans that carried their tend- 


ency to coil anteriorly to such a point that the resulting inconven- 
ience of a coil poised above the head gave survival value to any muta- 
tion that set the coil asymmetrically to one side, as it were? This 
again appears to be extremely unlikely. It would require a sym- 
metrically coiled predecessor and no such form is known. 

Finally then, I share the doubts of my predecessors, expressed or 
implied, that Pelagiella and its allies are gastropods^ but I shall go 
further than they have gone in that I shall not attempt to force them 
into a phylogenetic classification of the gastropods that appears to 
have no place for them. I shaU very tentatively assign them the 
peripheral position of a shoot from the same pre-Cambrian root as 
the main trunk of the gastropod family tree originating obscurely in 
pre-Cambrian and of otherwise little-understood affinities. It would 
be helpful if I could assign them elsewhere in the animal kingdom, 
but I cannot do so. 

Before leaving the subject of Pelagiella and its allies it may be 
well to review briefly and in general terms their chief characteristics. 
The shells are coiled and of from one-half to about three whorls. The 
coil is always asymmetrical. For the most part they are small and 
many are minute, a millimeter or two in diameter. The apical end of 
the whorl, the nucleus, appears in some to be laterally flattened and 
somewhat blunt, reminding one of the tip of a ram's horn, and in 
some forms slightly swollen. The whorls are ovoid in section, the 
narrow^ end of the ovoid being at the periphery. The spire is always 
low, varying in that respect from depressed to umboniform. In the 
forms w^ith a depressed spire the base is arched; in those with an 
arched spire it is flattish- The shells of any one species appear to 
be rather variable and it is probable that both dextral and sinistral 
forms occur in some species^ The ornamentation consists of fine, 
faint lines of growth and, on some forms, a single faint revolving 
lira, seemingly both above and below the periphery. The growth lines 
are somewhat drawn back at the rounded periphery, thus suggesting 
a broad, peripheral sinus. No operculum is known and there is no 
information on muscle scars. In some specimens of Pelagiella, 
Matthews reports and figures a groovelike constriction in the shell 
(or its steinkern ?) close to the apertural margin. It is not present on 
Matthew's primary types of Pelagiella atlantoides (Matthew), the 
genotype (Knight, 1941, p, 237), but does occur on rare specimens 
subsequently assigned to the species by Matthew. It may be a mark 
of maturity or old age. Possibly the constriction is seen only on the 
steinkern in which case it might mean only that the apertural margin 


was thickened within and the thickening is invisible on the outer 
surface of the shell. 

This group, in which I include Cambrian species mistakenly referred 
by authors to such genera as Straparollina Billings, Straparohis Mont- 
fort, Eiiomphahis Sowerby^ Raphistoma Hall, Ophileta Vanuxem, 
and Platyceras Conrad, is in urgent need of intensive study, as are 
all Cambrian gastropods and gastropodlike forms for that matter. 
Several names have been proposed for supposed genera, mostly on 
the basis of quite inadequate studies. Besides Pelagiella Matthew, 
1895, there are Par a pelagiella Kobayashi, 1939 (p. 287), ProtO' 
scaevogyra Kobayashi, 1939 (p. 286), and Proeccyliopterus Kobay- 
ashi, 1939 (p. 286). The last three seem to be erected on characters 
of very doubtful value or are differentiated from Pelagiella on mis- 
taken concepts of the characters of Pelagiella itself/* Still another 
genus, Semicircnlarea Lochman, in Lochman and Duncan, 1944 
(p. 44)^ was erected for the forms with only about one-half whorl 
often misidentified as Platyceras by previous authors. Pelagiella and 
its allies range throughout Cambrian and perhaps into early Ordovician 


As a result of our findings on our descent of the family tree and 
of the paleontological and neon to logical considerations given above, 
we have arrived at tentative hypotheses that force on our attention 
certain taxonomic conclusions. The first is that, since the mono- 
placophoran gastropods seemingly share with the polyplacopborans 
the basic isopleuran plan of organization, the two should be brought 
more closely together than has been customary in most classifications. 
The second is that, since the Anisopleura as the result of mutation 
arose suddenly from a monoplacophoran ancestor, and since certain 
anatomical features of both are very similar, the relationship between 
them is too close to permit them to be arranged in separate classes 
comparable in degree of differentiation to the other molluscan classes. 
The third is that, although the isopleurans and the anisopleurans 
should be placed in a single class^ the gulf betw^een them, both ana- 
tomically and in time, is profound and that, therefore, it seems ap- 
propriate to rank each as a subclass. On the basis of these three 
considerations we present the following revised definitions of the class 

^* It is unfortunate that the belief that the supposed characters of two of them 
suggested that they were ancestral to later genera of quite different affinities and 
led to the fixation of those ideas in the names given them< The proposal of 
names embodying phylogenetic concepts is most unwise. 


Gastropoda and its major subdivisions, the subclasses Isopleura and 
Anisopleura. There are also other condusions as to the subdivisions 
of the next lower rank but consideration of these is postponed until 
the class and the two proposed subclasses are dealt with. 

Class GASTROPODA Cuvier 

So great is the range of special morphological modifications in the 
class that it is exceedingly difhcult, if not impossible, to draw up a 
brief diagnosis that will cover all gastropods without excluding some 
forms that clearly must be included. The return of the isopleurans 
to the class, however necessary it appears, increases the difficulties, 
for we thereby reduce the convenient criteria of torsion and of a 
single shell to a status diagnostic of subdivisions of lower rank. 

The gastropods may be defined as mollusks with a differentiated 
head, a flat creeping foot, and a single basically conical shell. In a 
few gastropods specialized for free sw^imming the foot may be modi- 
fied into finlike organs, in the polyp lacophorans the primitive single 
shell has been divided transversely into eight segments, and in some 
highly specialized forms the shell has disappeared in the adult. In 
many others the cone is attenuated and coiled. Primitively marine, 
they have become adapted also to fresh waters and to terrestrial life. 
They are found at nearly all latitudes and nearly all altitudes from 
the depths of the oceans to high mountains. They appear in the fossil 
record in Lower Cambrian rocks and are flourishing today. 

Subclass Isopleura Lankester, — Gastropods that retain through- 
out life both in the shell and in the soft anatomy the primitive bi- 
lateral symmetry of the class. They are entirely marine and always 
rare- They first appear in the fossil record in Lower Cambrian rocks 
and carry through to the present* They probably originated in pre- 
Cambrian time. 

Subclass Anisopleura Lankester. — Gastropods that undergo tor- 
sion during the veliger stage. The Anisopleura are often abundant 
and are tremendously diversified in morphology and in habitat. They 
first appear in the fossil record as primitive forms in Lower Cam- 
brian rocks and are flourishing today, 


Order Polyplacophora, — Isopleuran gastropods with the shell 
made up of eight plates arranged along the midline of the dorsum; 
head not provided with eyes ; shell eyes, or aesthetes, may be present, 
Polyplacophora range from late Cambrian time to the present, are 
always marine and relatively rare. 


The subdivisions of the Polyplacophora will not be considered here- 
Order Monoplacophora. — Isopleuran gastropods with a single 
conical shell with the apex subcentral or pointed forward ; some pos- 
sibly with aesthetes. Marine, Lower Cambrian-Devonian. 

Before considering the subdivisions of the Monoplacophora it may 
be well to repeat that I do not consider Discinella Hall, 1871, Mober- 
gella Hedstrom, 1923, or Barella Hedstrom, 1930, to be monoplaco- 
phoran gastropods but hyolithoid opercula. Concho peltis Walcott, 
1879, I regard as probably a scyphozoan and certainly no moUusk, 
Chtiaria Walcott, 1899, is entirely problematical (Knight, 194 1, 
p. 20.) 

Family Tryblidiidae Pilsbry, 1899 

Subfamily PALAEACMAEINAE Grabau and Shimer, 1909 

Relatively low to high, cap-shaped shells with apex subcentral to 
slightly anterior. Muscle scars (observed only in Archaeophiala) 
discrete and arranged in six (or eight?) symmetrical pairs; orna- 
mentation basically concentric undulations. 


SceneUa Billings, 1872 (pi. i, fig, 1) {—ParmophoreUa Matthew, 

1886), Cambrian. 
HekioneUa Grabau and Shimer, igog (pL i, fig. 2), throughout the 

Palaeacmaea Hall and Whitfield, 1872, Upper Cambrian. 
Archaeophiala Koken, in Perner, 1903 (pi 1, fig* 3) (=Scaphe Hedstrom, 

1923, S capita Hedstrom, 1923, Patelliscapha Tomlinj 1929, and 

Paterella Hedstrom, 1930), Ordovician, 
Calloconus Perner, 1903, Lower Devonian, 

In the Silurian what appears to be a new genus hitherto unrecog- 
nized is represented by Palaeacmaea f solarium Lindstrom, 1884 

(p- 59)- 

Subfamily TRYBLIDHNAE Pilsbry, 1899 

Spoon -shaped shells with the apex at or overhanging the anterior 
end. Muscle scars (observed in TryhUdium, Pilina, Drahomira, Prop- 
Una, and partially in Cyrtonella) essentially similar to those of the 
foregoing family ; ornamentation concentric-lamellar or radiating. 


Tryblidium Lindstrom, 1880, Silurian, 
Cyrtonella Hall^ 1879, Devonian.^^ 

1^ I have given my reasons for including Cyrtonella Hall in the Tryblidiidae 
elsewhere (Knight, 1947b, p. 267), 


Helcionopsis Ulrich and Scofield, 1897, Ordovician. 
Drahomira Perner^ 1903, Ordovician.^^ 
Vallatoiheca Foerste, 1914 Ordovician, 
Pilina Koken, 1925, Silurian. 

Proplina Ulrich and Bridge in Kobayashi, 1933, Upper Cambrian- 
Lower Ordovician.*^ 

Family HypsELOCONiDAE, new 

Narrowly conical shells with the apex over the narrower (anterior ?) 
end but tilted slightly backward- Ornamentation growth lines or 
faint radiating undulations; muscle scars unknown. 


Hypsehconus Berkey» 1898, Upper Cambrian-Lower Ordovician,*^ 
Pollicina Holzapfel, 1895, Ordovician, 

Family Archinacellidae, new 

Low conical shells with the apex at or overhanging the anterior 
end. Ornamentation growth lines or radiating lirae; muscle scar a 
broad, continuous ring, narrowing in front where it passes below the 
apex, Ordovician. 


Archinacella Ulrich and Scofield^ 1897. 
tPiychopeltis Perner, igo3.i» 

It is possible that Helcionopsis will find a place here rather than 
with the Tryblidiidae when its muscle scars are discovered. 

Order Aplacophora, — In this order there is no shell, and it is 
consequently unknown as a fossil, I have no comments but retain it 

'^^ Drahomira is a name published, but not adopted, by Perner, 1903 (p. 23, 
footnote) for Tryblidium glaseri Barrande in Perner, 1903 (p. 23), genotype 
by monotypy. This name was overlooked by me in the preparation of 'Taleozoic 
Gastropod Genotypes" (Knight, 1941)- Seemingly it is the valid name for a 
distinct genus of this family. 

^^ The muscle scars of Proplina cornutaformis (Walcott), the genotype and 
only species referred to the genus in published literature, are unknown. How- 
ever, the material assembled for a monograph on Ozarkian and Canadian gas- 
tropods by E, O. Ulrich and Josiah Bridge is available to me and several 
species referred to the genus show them clearly. 

1* For comments on the supposed multiple paired muscle scars of Hypselo- 
conus see Knight, 1941 (p. 158), 

1^ Although Perner described a scar for Ptychopeltis, examination of his 
specimens failed to disclose valid evidence for it (Knight, 1941, p, 288). 


Subclass ANISOPLEURA Lankesteji 

Except for reviving Latikester's subclass Anisopleura, equivalent 
without the Monoplacophora (Tryblidiacea of Wenz) to the class 
Gastropoda of Wenz, 1938, I am now proposing few changes. To 
the Bellerophontacea, which are retained as Prosobranchia, are added 
the family Coreospiridae with the genera Coreospira Saito (pi. i, 
fig. 7), Cydohokus Knight, and Oelandia Westergard (pL i, fig. 5), 
but not without a residuum of doubt, and the superfamily Macluritacea 
is erected in the Prosobranchia* The Macluritidae of Wenz form its 
nucleus and other families composed of related elements are as- 
sembled with it. The revised taxonomy of the group will be presented 
as a part of another paper. The included genera are listed on pages 
36-37 of this paper. In all other respects the Anisopleura are left as 
Wenz had them but because of lack of opportunity for intensive study 
rather than because of detailed endorsement of his arrangements. 




The genus ''Bellerophon'' of the older workers and some neon- 
tologists (now expanded to a superfamily, the Bellerophontacea, with 
four families and something like fifty genera and subgenera) has 
been difficult to understand and to classify. Its isostrophic habit of 
coiling is almost unique in the Gastropoda. This and the fact that 
the entire superfamily has been extinct since Triassic time and affords 
no living examples from which soft parts can be demonstrated have 
seemingly left us with little information to go on. The broad mor- 
phological pattern of the soft parts must be inferred since it cannot 
be observed directly. 

De Koninck in 1883 (p. 121) reaffirmed on a more rational basis 
his suggestion of 1843 (p. 337) that the bellerophonts were proso- 
branch gastropods. Before 1883 the bellerophonts had been regarded 
as cephalopods, or as heteropod gastropods. Some specialized forms 
such as Pterotheca^ originally described as brachiopods or pelecypods, 
have been regarded as pteropods. Since that time they have been 
classified as prosobranch with the Docoglossa, or as a separate class 
of Mollusca, the ''Amphigastropoda;*' I can subscribe to none of 
these views except perhaps the main thesis of de Koninck in 1883, 
although not to the details. 

As stated previously, Wenz's great contribution to theory in 1938 


(p- 59) ^^s the idea that the Tryblidiacea were primitive untorted 
gastropods. This I applaud and accept. But he also regarded the 
bellerophonts as untorted gastropods similar to the Tryblidiacea, on 
the grounds of symmetry and an unsupported assumption that the 
slit and band are not to be compared with the seemingly homologous 
feature in the pleurotomarians. On this point I must part company 
with him. Curiously, if I read the story aright, Wenz seems to be 
following part way in the footsteps of many of his predecessors^ who 
regarded the bellerophonts as prosobranchs on the grounds of the 
following chain of reasoning: The early cup- shaped shells are sym- 
metrical and resemble the living patellids ; therefore they are to be 
classified with the latter as Docoglossa and prosobranchs. The bellero- 
phonts are also symmetrical ; therefore they are closely related to the 
patellids and are also Docoglossa and prosobranchs. 

But Wenz in recognizing the early cup-shaped mollusks, the Try- 
blidiacea, as nontorted gastropods changed the first premise of the 
customary chain of reasoning and the bellerophonts, still linked with 
these early cup-shaped shells, are, to Wenz, like them nontorted 

The weakness in both lines of argument is the overvaluing of the 
symmetry of bellerophonts as a criterion of relationship to the sym- 
metrical cup-shaped shells whether patellids or tryblidians, the under- 
valuing of the many manifest differences between the bellerophonts 
and either of the other two, and the undervaluing of several manifest 
anatomical homologies between the bellerophonts and the asymmetri- 
cal but coiled prosob ranch pleurotomarians. 

Fundamental to the undervaluing of bellerophont-pleurotomarian 
homologies is a failure on the part of Wenz and some neontologists 
to recognize that torsion and the development of lateral asymmetry 
are two distinct processes. Undoubtedly torsion set up unstable con- 
ditions that favored the natural selection of mutations, such as lateral 
asymmetry, that would result in a more efficient organism, but lateral 
asymmetry is not to be confused with torsion or what I have called 
torsional asymmetry. Although torsion is a prerequisite for asym- 
metry, asymmetry does not necessarily follow from it. It is as though 
this school of thought believes that the muscular pull that initiates 
torsion also distorts the lateral symmetry. I know no evidence that 
supports such a view. It is true^ of course, that except in the Isopleura 
all gastropods living today show lateral asymmetry at some ontoge- 
netic stage and the lateral asymmetry is initiated in the veliger larva 
immediately after torsion. But it does not follow that it was always 
universally thus. The view that asymmetry was the immediate or 


concommitant mechanical result of torsion and in consequence became 
a part of gastropod heritage smacks somewhat of Lamarckianism and 
in any case cannot be sustained. 

Returning to Wenz's views, in 1938 they seem to have been ap- 
proximately as I have stated them above. On the basis of the lateral 
symmetry alone he felt that the bellerophonts were closely related to 
the tryblidians and consequently had not undergone torsion. Al- 
though he gave no systematic expression to these views at that time, 
classifying both the TrybHdiacea and Bellerophontacea as Proso- 
branchia, it was his opinion that both were probably out of place in 
that position. 

Feeling insecure as to his interpretation of the bellerophonts he 
was quite rightly searching for corroborative evidence, and he felt 
that the discovery in the bellerophonts of multiple, paired dorsal 
muscle scars like those of the tryblidians would be strong supporting 
evidence, as indeed it would. In 1937 he wrote to me asking what I 
knew of bellerophont muscle scars and that started the chain of events 
about which I wrote ten years later (Knight^ 1947). Briefly, a speci- 
men of the supposed but somewhat atypical bellerophant Cyrtonella 
mitella (Hall) was discovered which seemed to support fully Wenz's 
views in that the unmistakable record of two pairs of dorsal muscle 
scars (not three as Wenz wrote) was clearly visible on that part of 
the steinkern that was exposed. Possibly other scars may be covered 
by matrix. Feeling that his views as to the close relationship of the 
bellerophonts and tryblidians were fully vindicated, Wenz published 
his paper giving systematic effect to those views by employing for 
them a subclass, the Amphigastropoda (Wenz, 1940).^** 

An interesting point about tryblidian muscle scars noted in Try- 
blidium, Archaeophiala, and Cyrtonella is that each scar has on the 
side tow^ard the margin of the shell a smaller, fainter scar as though 
it were the shadow of the scar cast before it. Wenz, who had never 
seen the specimen of Cyrtonella mitella he figured, misinterpreted a 
pair of these shadow scars, shown in the photograph sent him by Yang, 
as a principal scar* This is why he mistakenly reported three visible 
pairs of scars. The physiological significance of the '^shadow scars" is 

^** Actually the Amphigastropoda consisting of only the bellerophonts, was 
erected as a new class of mollusks by Simroth in 1906 (p, 839), who was fol- 
lowed by Thiele in 1935. Simroth*s course, and especially Thiele's, was sup- 
ported only by the gratuitous assumption that the soft anatomy was without 
torsion and bilaterally symmetrical. Amazingly, Thiele assumed also **eine 
anliche schwimmende Lebensweise — wie die Nautiliden*' (Thiele, 1935, p. 1125) 
in which he was followed by Wenz, 


obscure but their presence is an additional evidence that Cyrtonella is 
a tryblidian. 

Although the paired dorsal muscle scars on the specimen of 
Cyrtonella were discovered in my laboratory and although it was on 
my suggestion that Yang disclosed the discovery to Wenz, my views 
as to its significance were quite different from those so promptly pub- 
lished by Wenz in 1940. The more probable interpretation that 
Cyrtonella, a genus that was even then somewhat doubtfully placed 
in the Bellerophontacea and quite as easily interpreted as a tryblidian, 
should be placed in the Tryblidacea instead of the Bellerophontacea 
seems never to have occurred to Wenz. 

Fortunately I was able to discover the muscle scars of two unques- 
tionable bellerophont genera a few years later, Sinuites (pi i, fig. 11) 
and Bellerophon (ph i, fig, 13) (Knight, 1947). They consist of a 
single symmetrical pair. Each muscle was attached to the opposite end 
of the colummella about one-half whorl within the aperture, a posi- 
tion that would permit them to serve effectively as pedal retractors. 
They are not dorsal and not multiple pairs. Both of those facts are 
seemingly fatal to Wenz's arguments as to the closeness of the rela- 
tionship betw^een the tryblidians and the true bellerophonts. 

Wenz displays a number of views to which I must take exception. 
For example^ he accepts the wholly conjectural and long-rejected 
views of Lang (1891) as to the gradual development of torsion in 
the gastropods. He treats the bellerophonts and pleurotomarians as 
being present in early Cambrian rocks. In terms of genera recog- 
nized by him J neither appeared until late Cambrian time. Under the 
influence of his overestimate of the significance of external lateral 
symmetry in the bellerophonts he fails to even consider the close 
homologies betw^een bellerophonts and pleurotomarians. Finally, he 
seems to hold the view first proposed by Deshayes in 1830 (p. 135) 
and abandoned by most students well before the close of the nine- 
teenth century that the bellerophonts ''tended toward a freely swim- 
ming, nektonic mode of life" (translation from Wenz, 1938, p. 59)- 
I know of no evidence w^hatever that would support such a view and 
w^ould be interested indeed to learn of a molluscan swimming mecha- 
nism that W'Ould be powerful enough to sustain the massive shells 
of some bellerophonts above the sea bottoms. It seems highly prob- 
able that the bellerophont foot conformed in general to the pattern 
shown by other Archaeogastropoda. It was adapted to creeping, not 
to swimming, 

I regard the bellerophonts as prosobranch Archaeogastropoda close 
to and probably ancestral to the pleurotomarians which they precede 


ill the fossil record* The bellerophonts share with the pleurotomarians 
(i) a shell that typically has deeply hollow, usually closely coiled 
whorls, (2) a sinus or slit which, if a slit, generates a slit band, (3) a 
single pair of lateral retractor muscles, ^^ and (4) seemingly a single 
pair of each, of ctenidia, auricles, etc. They differ principally in that 
the coil of the bellerophont shell is bilaterally symmetrical (iso- 
strophic) and that of the pleurotomarian shell is an asymmetrical 
orthostrophic helicoid, in my view a difference of little significance 
for classification but of profound import for understanding gastropod 

Comparing the bellerophonts with the tryblidians we find they have 
one feature, and only one, in common : externally the shell of each is 
bilaterally symmetricaL But in respect to the first three categories 
in the foregoing paragraph, the tryblidians have (with a very few 
exceptionally high conical shells) (i) a shallow cup- or spoon-shaped 
shell with the apex bent toward one end, but no coiling, (2) no sinus 
or slit, and (3) multiple (usually six or eight) symmetrical pairs of 
dorsal muscle scars. 

Let us look for a moment at these points of agreement and dis- 
agreement. The agreement between the bellerophonts and pleuro- 
tomarians on points i and 3 can only mean that we have a shell with 
a deep body cavity into which the head and foot can be withdrawn 
by the retractor muscles which are properly placed in both for the 
operation. Point 2, the sinus or slit can only be an anal emargina- 
tion, a feature that is known otherwise only in prosobranchs and espe- 
cially in the Archaeogastropoda, and which is accepted by many 
neontologists such as Garstang, Yonge, Crofts, and many others as 
an adaptation for sanitation after torsion had created a need for it. The 
bilateral symmetry of the shell can no more be considered a charac- 
ter of subclass or even ordinal rank than that same symmetry can be 
employed to link two groups so different on other points as the 
bellerophonts and tryblidians. 

Need we continue to point further fundamental differences be- 
tween the pleurotomarians and bellerophonts on one hand and the 
trybhdians on the other, differences such as the impossibility of at 
least the low, cuplike tryblidians pulling their head and foot into the 
shell? In that respect they probably resembled the chitons and the 
secondarily symmetrical fissurellids and patellaceans- Since the muscle 

^^ The living pleurotomariid genera or subgenera Perofrochis Fischer, 
Entemnotrochus Fischer (pK 2^ fig. 4), and Mikadotrochus Lindholm, "Pleura- 
tomaria^' of authors, have only a single retractor muscle, although other living 
pleurotomarian genera have a pair. 


scars of the high and narrow species of Helcionella (pL i, fig. 4) 
in the tryblidians have never been observed, it is more difficult to 
speculate profitably as to whether the muscles are so placed that they 
could or could not have withdrawn the head and foot into the shelL 
It is possible that they could, especially if, as I am suggesting, the 
reduction of the pedal muscles from eight pairs to one pair may have 
occurred first in them. Need it be pointed out to those who regard 
the anal emargination in the bellerophonts as posterior that no known 
or reasonably imaginable nontorted mollusk has or needs a slit or 
sinus to provide egress for the contaminated water of a posterior 
anus? Certainly the chitons and the tryblidians do not. Again, have 
those who infer that bellerophonts are primitively orthoneurous non- 
torted "Amphigastropoda^* ever tried to imagine the animal with its 
large, heavy coil anterior and overhanging the head? The shell in 
many bellerophont species is not only thick and heavy but may carry 
a massive parietal callus as well. To me such an arrangement appears 
highly improbable, bordering indeed on the fantastic. It is suggested 
that the reader turn to figure 7, a, on page 30, where a restoration of 
a small very primitive bellerophont, with no parietal callus, is pre- 
sented as though it were primitively orthoneurous and exogastric, 
may help him to visualize it. 

In summary, it appears to me that the evidence for the view that 
the bellerophonts were prosobranchs, is very strong and the evidence 
that they were primitively orthoneurous '^Amphigastropoda" very 
weak indeed, 

Sojt anatomy of the bellerophonts. — We now know enough of bel- 
lerophont shell morphology and enough of the morphology of living 
examples of the obviously related pleurotomarians that we may specu- 
late with considerable safety on the general nature of bellerophont 
soft anatomy and perhaps even on its physiology and habits. 

One may be quite confident that they were aspidobranchs with a 
high degree of bilateral symmetry reflected in symmetrically paired 
ctenidia, osphradia, hypobranchial glands, auricles, kidneys, and per- 
haps even gonads. They probably crawled on the sea bottom on a 
generalized gastropod foot- It seems probable that like other aspido- 
branchs they fed chiefly on vegetable matter and were rhipidoglossate. 
Nothing is known of the bellerophont operculum, if there was one. 

Perhaps a diagrammatic restoration q^^ some of the more significant 
soft parts with an interpretation of the course of the water currents 
in the mantle cavity will save pages of words. 

On figure 10 is shown a shell of a bellerophont species, Knightites 
multicorniitus Moore, 1941 (p. 153), with the soft parts restored in 

Arrows shew mantle cavity 


Functional inholant conols 


X^^'f A^^^ Slit (e 

Left osphrodium 

Left ctenidium 


ight osphrodium 

Right ctenidium 

/_ gijj^ band 

Anterior protuberance of the parietal callus. 
Pedal retractor musculature not shown. 


Abarvdoned inholonf conols 

Osphrodium ^^^ 
Slit (exh^lont) \ Rectum -^ ' 

Left ctenidium 

Area of ottochment of left 
pedal retractor muscle. 


Porietal Collus 

Footjj with left pedal retractor muscle^ 
passing on the near side of the parietal 



Figure io 
(See opposite page for explanation.) 



terms of the above interpretation. It is a modification of figure yd in 
Moore, 1941 (p. 15S). I have abandoned the ideas expressed in 
figures 7a-c as untenable in the light of more accurate knowledge of 
the aerating currents in Haliotis than I then had. The extended 
periodic, paired canals on each side of the slit and slit band in K. nmlti' 
cormttus interested me very much. It occurred to me that they gave 
a clue to the region on the mantle lip through which passed the cur- 
rents of water that aerated and flushed out the mantle cavity* The 
works of Yonge and Crofts on the aerating currents in various gas- 
tropods including the pleurotomarian Haliotis seem to reinforce the 
suggestion made by the canals of K. multicornutits, so that one can 
infer the probable course of the principal water currents in that species 
and probably in all generaHzed bellerophonts. This inferred circula- 
tion is in all respects that of a prosobranch and seems a reasonable 
approximation to the probable condition during life. 

Crofts, D. R, 

1937. The development of^ f/a/fo/u tuberculafa, with special reference to 
organogenesis during torsion. Philos, Trans. Roy. Soc. London, 
sen B, No, 552, vol 228^ pp. 219-268. 

Garstang, W. 

1929, The origin and evolution of larval forms* British Assoc* Adv. Sci., 
Rep, g6th Meeting, pp. 77-98, 

Howell, B. F,, et al, 

1944, Correlation of the Cambrian formations of North America* Bull, 
GeoL Soc. Amer., vol. 55, pp. 993-1004. 

Knight, J. Brookes. 

1941, Paleozoic gastropod genotypes, Geol, Soc, Amer. Spec. Pap. 32- 
1947a, Some new Cambrian bellerophont gastropods, Smithsonian Misc, 
Coll., vol, 106, No, 17, 

Figure 10 

Two views of a diagrammatic restoration of Knighiites mulikornuHis Moore 
(Bellerophontacea, Bellerophontidae), approximately X 2. The shell is drawn 
as though it were partly transparent so as to show some of the fleshy organs. 
The arrows show the probable path of the principal aerating and cleansing 
currents. Although most bellerophonts did not have inhalant canals as did 
Knightitcs it is thought that the path of the currents and regions of their 
entrance and exit were approximately the same as inferred for Knightites. 
In living Haliotis where the details are known the paths are homologous in 
every respect 

a, Dorsal view. 

frj Left side view. The fibers of the left retractor muscle are shown in a 
highly schematic fashion as though they anastomose into the muscles of the 
foot (which are not shown). 


1947b* Eellerophont muscle scars, Journ. PaleontoL, vol. 21, pp. 264-267. 
1948, Further new Cambrian bellerophont gastropods, Smithsonian Misc, 
ColL, vol* III, No. 3. 


1939* Restudy of Lorenz's Rapbistoma hroggeri from Shantung, with a 
note on Pelogiella. Jubilee publication in commemoration of Pro! 
H, Yabe's birthday, 
LankesteRp Ray, 

1S78, Mollusca, Encyl. Britt., 9th td., vol 16, 


1884. On the Silurian Gastropoda and Pteropoda of Gotland. Kong: 

Svenska Vet.-Akad. HandL, vol. 19, No. 6, 
LocHMAN, Christina. 

1944, In Lochman and Duncan, Early Upper Cambrian faunas of central 

Montana. Geoh Soc. Amer. Spec. Pap. 54. 

Moore, R. C. 

1 941- Upper Pennsylvanian gastropods from Kansas, State Geol, Surv. 
Kansas Bull. 38. 
Naef, a. 

191 L Studien zur generellen Morphologie der Mollusken, Teil i. t)ber 
Torsion und Asymmetrie des Gastropoden, Ergebn. und Fortschcr. 
Zool, vol 3, PP> 73-164* 
Pelseneer, P. 

1906. Moilusca, In Lankester, E. R,, A Treatise on Zoology, pt, 5. 


1906. In Bronn, H. G., Klassen und Ordnungen des Tier-Reich s, 
Thiele, J, 

1 93 1 -1 935, Handbuch der systemtischen Weichtierkunde. 
Wenz, W, 

1938-1944. Gastropoda. In Schindewolf, O. H., Handb. Palaeozool., voL 6, 

pt I, i"7, 
1940. Ur sprung und friihe Stammes-geschichte der Gastropoden. Arch. 
Molluskunde, vol. 72, pp. i-io. 
Woodward, S. P, 

1851-1856. A manual of the Mollusca, or rudimentary treatise of Recent 
and fossil shells. 


1939. On the mantle cavity and its contained organs in the Loricata 
(Placophora)* Quart. Journ. Micr. Sci., vol. 81, pp. 367-390. 

1947, The pallial organs in the aspidobranch Gastropoda and their evolu- 
tion throughout the Mollusca. Pliilos» Trans. Roy. Soc* London, 
ser* B, vol. 232, pp. 443-518. 

Just as this work reached page-proof stage, a copy of the "Traite de Paleon- 
tologie/* published under the direction of Prof. Jean Piveteau, reached Wash- 
ington (Traite de Paleoiitologie, vol. 2, 1952). The chapter on the Gastropoda 
is by Dr. Genevieve Termier (nee Delpey) and Prof. Henri Termier of Algiers, 
Since I have discussed Mme. Termier's view^s on gastropods elsewhere (Geol, 
Mag., voL 83, pp. 280-284^ 1946), I shall say nothing further here except to 
reaffirm my almost complete disagreement, — J. B. K.